an overview of nature-inspired, conventional, and hybrid...

Review ArticleAn Overview of Nature-Inspired Conventional and HybridMethods of Autonomous Vehicle Path Planning

Ben Beklisi Kwame Ayawli 12 Ryad Chellali34

Albert Yaw Appiah15 and Frimpong Kyeremeh15

1College of Electrical Engineering and Control Science Nanjing Tech University China2Computer Science Department Sunyani Technical University Sunyani Ghana3Nanjing Forestry University Nanjing China4Kita Technologies Nanjing China5Electrical and Electronic Engineering Department Sunyani Technical University Sunyani Ghana

Correspondence should be addressed to Ben Beklisi Kwame Ayawli bbkayawliyahoocom

Received 16 March 2018 Accepted 28 June 2018 Published 19 July 2018

Academic Editor Jose E Naranjo

Copyright copy 2018 Ben Beklisi Kwame Ayawli et al This is an open access article distributed under the Creative CommonsAttribution License which permits unrestricted use distribution and reproduction in any medium provided the original work isproperly cited

Safe and smooth mobile robot navigation through cluttered environment from the initial position to goal with optimal path isrequired to achieve intelligent autonomous ground vehicles There are countless research contributions from researchers aimingat finding solution to autonomous mobile robot path planning problems This paper presents an overview of nature-inspiredconventional and hybrid path planning strategies employed by researchers over the years for mobile robot path planning problemThe main strengths and challenges of path planning methods employed by researchers were identified and discussed Futuredirections for path planning research is given The results of this paper can significantly enhance how effective path planningmethods could be employed and implemented to achieve real-time intelligent autonomous ground vehicles

1 Introduction

In recent years mobile robot path planning research aimedat achieving intelligent autonomous ground vehicles hasreceived more attention Path planning and motion controlare very important but complex navigation tasks in roboticsPath planning in mobile robotics refers to strategies todetermine how a mobile robot gets to its goal safely ensuringthat obstacles are avoided [1] To successfully achieve pathplanning and motion control a vehicle is expected to possessthe ability to perceive and detect obstacles to be avoided toenable it to reach its target safelyWhat is expected in roboticsis to develop real-time intelligent autonomous navigationrobots that can sense and interpret information collectedfrom their environment to determine their position directionto goal avoid obstacles and perform smooth navigation inboth structured and unstructured environments The robotsare expected to perform these tasks with safe and shortest

path and lowest time to their target and finally performtheir assigned task without the intervention of humans [2ndash7] There are many applications and benefits to be derivedfrom intelligent autonomous robots Some of which includerescue operations during disaster task performance in facto-ries homes in areas of transportation medicine educationagriculture and many others

Mobile robotics path planning is described as multi-objective optimization problem since it is required to generatethe appropriate trajectories as well as obstacle avoidance inthe environment [8] Three tasks are required during mobilerobot path planning These include acquiring informationfrom its environment localization of its current positionand finally taking necessary decision based on the providedalgorithm and the information acquired to execute its task[9 10] Lavalle [11] identified feasibility and optimality aspath planning criterion with different concerns The con-cern of feasibility criteria is to determine a plan to reach

HindawiJournal of Advanced TransportationVolume 2018 Article ID 8269698 27 pageshttpsdoiorg10115520188269698

2 Journal of Advanced Transportation

goal irrespective of the efficiency Optimality is concernedwith determining an optimized feasible plan to reach goalwith efficient performance Comparing the two achieving afeasible path planning was described to be a problem butthat of optimal path planning is more problematic Pathplanning and obstacle avoidance methods could be classifiedinto static or dynamic based on the environment and asglobal or local based on the path planning algorithm [12]Environment composed of stationary objects is described asstatic while thosewithmoving objects are termed as dynamicWhen the path planning occurs in a static environmentand all the information of the environment required bythe robot to perform path planning and obstacle avoidanceis available it is described as global path planning Localpath planning on the other hand occurs when the robotis in motion and it reacts to changes in the environmentto decide its movement [1 10] Path planning can also bedescribed as either online or off-line and reactive or mapbased [13] Off-line techniques compute the path beforenavigation while online techniques perform the computa-tions and take decision incrementally during the navigationprocess Data derived from cameras and sensors includingultrasonic sensors infrared sensors and light sensors areinterpreted in different ways by algorithms to embark on safepath planning Control of navigation and obstacle avoidanceformobile robots can be described as a reactive approach [14ndash16] because the robot is expected to react against its changingenvironment by reacting immediately to new informationreceived from sensors Control of path planning is howeverdescribed as deliberative approach because the robot isexpected to provide the exact planning to achieve a goalrelying on the geometric model of the environment and theappropriate theory The most common control method ofavoiding obstacles in cluttered environment by mobile robotis the reactive local navigation schemes where the robotrsquosaction depends on sensor information

Ongoing research in mobile robotics is geared towardsbuilding autonomous and intelligent robots that performrobust motion planning and navigation in dynamic envi-ronments [17] There are records of considerable number ofresearch that aimed at addressing path planning and obstacleavoidance problem using diverse approaches and algorithms[17ndash29] Despite the numerous efforts to address the problemof safe path planning of mobile robots there still exists achallenge in achieving real safe and optimal path planningto achieve the use of intelligent autonomous vehicles [5 30ndash35] These outstanding challenges motivated this research toconduct an overview of the popular methods and strategiesused by researchers to address the problem and to identifythe strengths and challenges of these approaches and con-sider the way forward to achieving intelligent autonomousvehicles

Mobile robot path planning methods could be catego-rized in different ways In this paper we classified theminto nature-inspired computation conventional and hybridmethods Methods and strategies which have something todo with the imitation of phenomena of nature are describedas nature-inspired computation method Those that are notrelated to imitation of nature phenomena are classified under

conventional method Methods that combine two or morestrategies are described as hybrid methods in this paper

The remainder of this paper is organized as follows dis-cussion of conventional mobile robot path planningmethodsis given in Section 2 The strengths and challenges of theconventional methods are given in Section 3 Discussion ofnature-inspired mobile robot path planning methods withtheir strengths and challenges is covered in Sections 4 and 5In Sections 6 and 7 hybrid methods with their strengths andchallenges are presented Conclusion and the way forward formobile robot path planning and obstacle avoidance are givenin Section 8

2 Conventional Path Planning Methods

The conventional path planning methods (CPPM) are thetraditional methods used over the years by researchers formobile robot path planning Most of these methods rely ondistance information from objects to the robots force ofattraction and repulsion statistical methods clustering orgraphical map calculations to determine the path planning ofthe robotThey are comparatively computationally expensiveNotable among them are artificial potential field (APF)distance control algorithm bumper event approach wall-following sliding mode control (SMC) dijkstra Alowast DlowastStereo block matching voronoi diagram (VD) simultaneouslocalization and mapping (SLAM) vector field histogram(VFH) rapidly exploring random tree (RRT) curvaturevelocity lane curvature dynamic window and tangent graphAccounts of the use of some of these approaches are given inthe next subsections

21 Artificial Potential Field Path Planning Method One ofthe popular methods used by researchers over the years isAPF APF is a mathematical method which causes a robotto be attracted to the goal but repelled by obstacles in theenvironment [6 34 36 37] There are notable research workscarried out using APF with sensors to plan the path of mobilerobots to provide autonomous navigation taking obstacleavoidance into consideration [12 37ndash47]The idea of applyingAPF to path planning originated from Khatib [34] APFapproach was modified by some researchers to make it moreefficient for path planning and obstacle avoidance [41 4348 49] The algorithm of APF comprises the attractive forcerepulsive potential force and the total potential force APFalgorithmusingGaussian function as in [50] can be expressedas shown in the following

119860119865 (119901) = 119891119886119905119905 [1 minus 119864119883119875 (minus119888119886119905119905 lowast 1198892119892)] (1)

119877119865 (119901)

= sum

119891119903119890119901119894 [119864119883119875 (minus119888119903119890119901 lowast 1198892119900119887119904119894)] 119889119900119887119904119894 (119901) le 11988900 119890119897119904119890

(2)

119865119905119900119905119886119897 (119901) = 119860119865 (119901) + 119877119865 (119901) (3)

where AF(p) is total attraction force fatt is maximum value ofattraction force at any instance catt is attractive constant dg is

Journal of Advanced Transportation 3

Euclidean distance between the robot and the goal RF(p) istotal repulsive force frep is maximum value of repulsive forceat any instance crep is repulsive constant dobs is Euclideandistance between the robot and the obstacle

One of the recent usages of APF in mobile robot pathplanning is found in [40] They extended the APF methodby considering household animalsrsquo motion attributes withthe purpose of improving human-robot interaction Theattractive and the repulsive forces of APF were all extendedA real robot experiment was conducted using holonomicrobot with six cameras to test their algorithm and to confirmthe simulation results Also to help solve the point massproblem of APF for car-like robots potential field windowwhich is an extension of APF method was proposed in [51]to provide safe navigation With this method the potentialfield computations were done differently compared to theconventional APF method APF has also been considered inmultiple-robot path planning in [52]

22 Vision-Based Path Planning Method Although mostapproaches use information from cameras and sensors todetermine the execution of their algorithms some methodsrely basically on image processing of information from cam-erasThese methods were described as vision-based methods[33 53ndash57] A recent visual-based approach was presentedin [56] to deal with path planning and obstacle avoidanceproblem for small unmanned vehicles by adopting the regioninterest extraction method used in [57] Local blind decon-volution was utilized to classify the regions of the imagescollected relative to each other to help produce a feature mapcomposed of localized structural formation of the processedimages The feature maps realized were then used as the basisfor the detection and obstacle avoidance This approach wasfirst introduced in [58] With this approach before the robotdecides to move to a given direction images are capturedand downsized to 320-pixel column width to aid fastercomputations The feature map is then extracted The partsof the map are then used to determine if there are obstaclesbefore the robot makes a move to the appropriate directionComparative results demonstrated that this method resultedin less frequent obstacle hits of 4-5 compared to histogram-based contrast (HC) region-based contrast (RC) and spectraresidual (SR) which registered between 11 and 14 hits

A visual-based navigation algorithm for mobile robotusing obstacle flow extracted from captured images to deter-mine the estimated depth and the time to collision usingthe control law called balanced strategy was also presentedin [53] A combination of very low-resolution images and asonar sensor was also used to develop a vision-based obstacledetection algorithm in unstructured indoor environment in[33] While a digital camera was used to take images forsegmentation the sonar sensor was used to extract depthinformation from the image Kim and Do [54] on the otherhand presented a visual-based dynamic obstacle detectionand avoidance approach with block-basedmotion estimation(BBME) using a single camera A single camera sensor basedon relative focus maps method was also presented in [58]Extracted region of images was taken and divided into 3x3regions to determine the intensity of the regions The region

with high intensity specifies an obstacle and those with lowintensity determine the way to go during navigation

Instead of using cameras as others did Lenser and Veloso[55] demonstrated visual sonar method to detect known andunknown objects by considering the occlusions of the floor ofknown colors to aid detection and obstacle avoidance Withthe assumption that target velocity is known remaining thatof the speeds of dynamic obstacles a sensor-based onlinemethod termed as directive circle (DC) for motion planningand obstacle avoidance proposal wasmade in [59] Both staticand moving obstacles were considered in a simulation usingthe approach

Recently a visual navigation approach based on transferlearning was presented in [60] to enhance the environmentalperception capacity in semantic navigation of autonomousmobile robots The method is made up of three-layer mod-els including place recognition rotation region and siderecognition Results from real experiments indicate goodperformance of themethod in semantic navigationThe robotwas able to recognize its initial state and poses and performedpose correction in real time Improvement is required for itsimplementation in complex and outdoor environment

23 Wall-Following Path Planning Method Wall-following(WF) method was also considered by researchers WFmethod considers the wall around the robot to guide itmovement from one location to another by the help of rangesensors Gavrilut et al [61] demonstrated a wall-followingapproach to mobile robot obstacle avoidance by consideringthe faster response time and easy integration of IR sensorsinto microcontrollers Robby RP5 robot equipped with twoIR proximity sensors was used to test their algorithmThoughthe completion time during the experiment was low errorswere generated because of interference of emitted signalsfrom the two sensors used In addition relatively smallobstacles could not be detected

24 Sliding Mode Control Path Planning Method In a dif-ferent development sliding mode control (SMC) was con-sidered by some researchers [62ndash64] SMC is a nonlinearcontrol method that brings about changes in the dynamics ofa nonlinear system by applying discontinuous control signalthat force the system to slide towards a cross section of thenormal behavior of the system Matveev et al [62] presenteda slidingmode strategywhich ismathematically expensive forborder patrolling and obstacle avoidance involving obstaclesinmotion Obstacles were considered to have different shapesrandomly at different periods The velocities of the movingobjects were considered to aid in the avoidance of obstacles inmotion Simulation and experiment with unicycle-like robotwere performed to evaluate the approach

25 A Reactive Dynamic Path Planning Method A reactivedynamic approach employs sensor-based or vision-basedapproach by reacting to unforeseen obstacles and situationsduring navigation with appropriate decisions A reactivedynamic approach to path planning and obstacle avoidancehas been implemented in autonomous mobile robots overthe years [12 65ndash67] One of such works is seen in Tang and


Ang [66] Based on situated-activity paradigm and divide-and-conquer strategy they adopted a reactive-navigationapproach to propose a method called virtual semicircles(VSC) which put together division evaluation decision andmotion generation modules to enable mobile robots to avoidobstacles in complex environments Simulation was usedin the implementation of this approach Matveev Hoy andSavkin [5] proposed a computational expensive approachtermed as a reactive strategy for navigation of mobilerobots in dynamic environment unknown to the robots withcluttered moving and deformed obstacles This approachwas implemented using simulation A reactive-navigationapproach using integrated environment representation wasproposed for obstacle avoidance in dynamic environmentwith variety of obstacles including stationary and movingobjects in [68] The approach was experimented on a PioneerP3-DX mobile wheeled robot in an indoor environmentA method using reactive elliptic trajectories with reactiveobstacle avoidance algorithm embedded in a multicontrollerarchitecture to aid obstacle avoidance was also presented in[69]

26 Dynamic Window Path Planning Approach Someresearchers also considered the dynamic window approach(DWA) to provide optimal obstacle avoidance [70ndash74]Recently an improved dynamic window approach wasproposed in [75] for mobile robot obstacle avoidanceConsidering the drawbacks of local minima of this approachcausing the robot to be trapped in a U-shaped obstacle alaser range finder was used as the sensor to ensure optimapath decision making of the robot The size of the robot wasconsidered in this approach Simulation was used to evaluatethe approach and compared with other DWA results inMATLAB and the former performs better according to theauthors

27 Other Conventional Path Planning Methods Accountsof other conventional path planning methods used byresearchers other than those described above are summarizedbelow

A heuristic A-Star (Alowast) algorithm and dynamic steeringalgorithm were employed in [10] to propose an approach toobtain the shortest path and the ability to avoid obstaclesthat are known to the robot in a predefined grid-basedenvironment map based on information derived from sonarsensors

A research on using robot motion strategies to enable arobot to track a target in motion in a dynamic environmentwas done in [76] Four ultrasonic range sensors with twocontrol algorithms where one controlled the stopping of thevehicle upon sensing an obstacle and the other deciding thedirection of movement were presented in [77] This methodwas implemented on a three-wheeled mobile robot in indoorenvironment

Distance control algorithm for mobile robot navigationusing range sensors was also considered by researchersfor path planning Notable among them was the researchby Ullah et al [78 79] They employed distance controlalgorithm to develop a remote-controlled robot to predict

Figure 1 Images showing the vehicle moving towards goal whileavoiding obstacles using SGBM (source [83])

and avoid collisions between vehicles by maintaining a givensafe distance between the robot and the obstacle

In another development a bumper event approach wasemployed to develop an algorithm for obstacle avoidance ofTurtlebot in [80] The approach however did not providecollision free navigation

Kunwar et al [81] conducted a research into usingrendezvous guidance approach to track moving targets ofrobots in a dynamic environment made up of stationaryand moving obstacles To enable efficient noise filtration ofdata collected from the environment Chih-HungWei-Zhunand Shing-Tai [82] recently evaluated the performance ofExtended Kalman Filtering (EKF) and Kalman Filtering (KF)for obstacle avoidance for mobile robots Implementationwas carried out using a two-wheeled mobile robot withthree sonar sensors EKF approach was described to haveperformed better than KF in the experiment

Semi-global stereo method (SGBM) with local blockmatching algorithm (BM) where obstacles were identifiedusing a method that checks the relative slopes and heightdifferences of objects was used in [83] A disparity mapobtained from the processed pair of images using stereocameras was taken through further processing and finally tothe obstacle avoidance algorithm to determine themovementof the vehicle to the defined GPS point while avoidingobstacles Real-time experiment was performed using anelectric vehicle as shown in Figure 1

Other methods including velocity space [84] SLAMVD [85 86] VFH [87] RRT [88 89] and others have alsogained popularity in the field of mobile robot path planningRecently a visual SLAM system-based method was proposedin [90] for a mobile robot equipped with depth sensor orcamera to map and operate in unknown 3D structure Thepurpose of the approach was to improve the performanceof mapping in a 3D structure with unknown environmentwhere positioning systems are lacking Simulation was done


using Gazebo simulator Real experiment was performedusing Husky robot equipped with Kinect v2 depth sensor andlaser range scanner Results confirmed improvement againstclassical frontier-based exploration algorithms in a housestructure Improvement may be required to address how todifferentiate similar structures and to deal with errors indetermining waypoints

3 Strengths and Challenges of ConventionalPath Planning Methods

Theconventional path planningmethods have their strengthsand weaknesses These are discussed in the next subsections

31 Strengths of Conventional Path Planning Methods Eventhough conventional methods of path planning are compu-tational expensive they are easy to implement Methods likeAPF DWA Alowast PRM and other conventional methods aresimple to implement The implementation of SMC strategyfor instance is easy and fast with respect to response timeIt also performs well in a condition with uncertain andunconducive external factors [64] Also these methods cancombine well with other methods and they perform betterwhen blended with other methods Popular among them isAPF which many researchers combined with other methodsAlowast for instance is very good at obtaining the shortest pathfrom the initial state to the goal When combined with othermethods the hybrid method is able to generate optimal path

32 Challenges of Conventional Path Planning MethodsNotwithstanding the strengths of conventional methods forpath planning there exist challenges affecting the achieve-ment of intelligent autonomous ground vehicles The follow-ing are some of the challenges with these methods

To begin with most of the approaches including con-ventional methods discussed used cameras and sensors tocollect data from the environment to determine the exe-cution of the respective path planning algorithms [55 91ndash93] Unfortunately readings from cameras and sensors areinterfered with noise due to changes in pressure lighteningsystem temperature and others This makes the collecteddata uncertain to enable control algorithms to achieve safeand optimal path planning [92 94] The dynamics of thevehicles also cause noise including electric noise of themobilerobots [5] These conditions affect the accuracy efficiencyand reliability of the acquired real-time environmental datacollected to enable the robot to take a decision [95] Researchwork has been done to address the noise problem but it is stilla challengeThis has effect on the practical implementation ofproposed approaches

Typical with vision-based obstacle avoidance strategiesfactors including distance of objects from the robot colorand reflection from objects affects the performance of detect-ing obstacles especially moving objects [54]The use of stereovision approaches [96 97] is very limited and can work onlywithin the coverage of the stereo cameras and could not workin regions that are texture-less and reflective [98] There is aproblem identifying pairs of matched points such that each

of these points demonstrate the projection of the same 3Dpoint [99] This results in ambiguity of information betweenpoints of the images obtained which may lead to inconsistentinterpretation of the scene [100]

Furthermore some of the approaches rely on the environ-mental map to enable the robot take decisions in navigationThere is a problem with unnecessary stopping of the mobilerobot during navigation to update its environmental mapThis affects the efficiency and the real applications of mobilerobots to provide safe and smooth navigation This challengewas considered in [101] and was demonstrated through simu-lation However it could not achieve the safe optimal path tothe target andwas not also demonstrated in a real experimentto determine its efficiency in support of simulation results

DWA also has the drawback of resulting in local minimaproblems and nonoptimalmotion decision due to constraintsof the mobile robot the approach could not manage [75 102ndash104]

In addition despite the popularity of the use of APFapproach in path planning and obstacle avoidance of mobilerobots it has serious challenges Navigation control of robotsusing APF considers the attraction forces from target andthe repulsion forces from the obstacles When performingthis attraction and repulsion task environmental informationis compared to a virtual force Computations lead to lossof important information regarding the obstacles and causelocal minima problem [36 40 41 105 106] APF can alsoresult in unstable state of the robot where the robot findsitself in a very small space [107] It can cause the robot tobe trapped at a position rather than the goal [51] Otherproblems of APF include inability of the robot to passbetween closely spaced obstacles oscillations at the point ofobstacles and oscillations when the passage is very small forthe robot to navigate and goals nonreachable with obstaclenearby (GNRON) It also performed poorly in environmentwith different shapes of obstacles [108ndash110] Limited sensingability bounded curvature limited steering angle and veloc-ity or finite angular and linear acceleration of the mobilerobot hardware have great effect on the performance of APFmethods [51] It may also be far from the optimum whenplanning is local and reactive [111]

Though VFH unlike APF strategy does well in narrowspaces if the size and kinematics of the robot are not takeninto consideration it can fail especially in narrow passages[112] Incorrect target positioning may also result in VFHalgorithm not being reliable [113]

Besides SMC methods are fast with respect to responsetime and are also good transient robust when it comes touncertain systems and other external factors that are notconducive [64] But it does not perform well if the longitu-dinal velocity of the robot is fast It also has the problem ofchattering which may result in low control accuracy [114]

Moreover some methods like VD RRT and others dowell in simulation environment but relatively difficult toimplement in real robot platform experiments VD RRTand PRM are good at generating global roadmap but a localplanner is required to generate the path They are not goodat performing in dynamic environments if not combinedwith other methods Alowast method on the other hand has the


problem with generating smooth path and it requires pathsmoothening algorithms to achieve smooth path

4 Nature-InspiredComputation-Based Methods

According to Siddique and Adeli [115] nature-inspired com-puting is made up of metaheuristic algorithm that try toimitate or base on some phenomena of nature given bynatural sciences A considerable number of researchers triedaddressing the problem of mobile robotics path planningand obstacle avoidance using stochastic optimization algo-rithm techniques that imitate the behavior of some livingthings including bees fish birds ants flies and cats [116ndash122] These algorithms are referred to as nature-inspiredparadigms and has been applied in engineering to solveresearch problems [123 124] Nature-inspired computation-basedmethods use ideas obtained fromobserving hownaturebehaves in different ways to solve complex problems thatare characterized with imprecision uncertainty and partialtruth to achieve practical and robust solution at a lowercost [125] Notable among nature-inspired methods usedin path planning and obstacle avoidance research includeartificial neural networks (ANN) genetic algorithms (GA)simulated annealing (SA) ant colony optimization (ACO)particle swarm optimization (PSO) fuzzy logic (FL) artificialbee colony (ABC) and human navigation dynamics Nature-inspired computation-based methods were claimed to bebetter navigation approaches compared to conventional onessuch as APF methods [126] Most of these methods considerreinforcement learning to ensure mobile robots perform wellin unknown and unstructured environments Accounts of theuse of some of these approaches are discussed in the nextsubsections

41 Fuzzy Logic Path Planning Method FL is one of themost significant methods used over the years for mobilerobot path planning Though FL had been studied since 1920[127] the term was introduced in 1965 by Lotfi Zadeh whenhe proposed fuzzy set theory [128] FL is a form of many-valued logic with truth values of variables of real numberbetween 0 and 1 used in handling problems of partial truthIt is believed that Fuzzy controllers based on FL possess theability to make inferences even in uncertain scenarios [129]FL can extract heuristic knowledge of human expertise andimitate the control logic of human It has the ldquoif-thenrdquo human-like rules of thinking This characteristic has made FL andother derived approaches based on FL become the most usedapproach by many researchers in mobile robot path planning[4 17 18 28 130ndash151]

Recently fuzzy logic was employed in [152] to proposemobile robot path planning in environment composed ofstatic and dynamic obstacles Eight sensors were used to col-lect data from the environment to aid detecting the obstaclesand determining trajectory planning Implementation wasdone in simulation and real-time experiments in a partiallyknown environment and the performance of the proposedapproach was described to be good Multivalued logic frame-work was used to propose a preference based fuzzy behavior

system to control the navigation of mobile robots in clutteredenvironment [147] Experiment was done using a Pioneer2 robot with laser range finder and localization sensors inindoor modelled forest environment The possibility of thisapproach performing in the real outdoor environment withconsiderable number of constraints and moving obstacles isyet to be determined

In another development FL based path planning algo-rithm using fuzzy logic was presented by Li et al [139]The positions of the obstacles and the angle between therobot and the goal positions were considered as the inputvariables processed using fuzzy control system to determinethe appropriatemovement of the robot to avoid collidingwithobstacles The approach was implemented in simulation

Fuzzy logic approach was employed in [153] to developa controller with 256 fuzzy logic rules with environmentaldata collected using IR sensors Webot Pro and MATLABwere used for the software development and simulationrespectivelyThe performance in obstacle avoidance using themethod in simulation was described to be good Howeversince no real experiment is implemented we could hardlyjudge if the approach could work in the real environmentwhere constraints including noise and the limitation ofhardware components of mobile robots are common Anapproach involving fuzzy logic to formulate the map of aninput to output to determine a decision described as fuzzyinference system (FIS) was employed to propose a methodto control autonomous ground vehicles in unstructuredenvironment using sensor-based navigation technique [28154] During robot navigation environmental mapping isdone inmost cases to enable the robot to decide itsmovement[147] At the course of navigation the robot stops to updatethe environmental map to determine the next move [14]However Baldoni Yang and Kim [101] are of the viewthat periodic stopping to update the environmental mapcompromises the efficiency and practical applications of theserobots Hence they proposed a model using a graphical andmathematical modelling tool called Fuzzy Petri net with amemory module and ultrasonic sensors to control a mobilerobot to provide dynamic and continuous obstacle avoidanceThe approach was demonstrated using simulation whichachieved 8 longer than the ideal path as against 17 longerthan ideal path with the approach without memory

Apart from the discussion above some researchers alsoemployed FL with the help of ultrasonic and infrared sensorsto develop path planning algorithms which were successfullyimplemented in both simulation and real robots platform [1725 135 144ndash151]

42 Artificial Neural Network Path Planning MethodResearch involving the use of ANN as intelligent strategyin solving navigation and obstacle avoidance problems inmobile robotics has also been explored extensively [102 155ndash166] The application of ANN tries to imitate the humanbrain to provide intelligent path planning

Chi and Lee [157] proposed neural network controlstrategy with backpropagationmodel to control mobile robotto navigate through obstacles without collision P3DXmobilerobot was used to evaluate the developed algorithm for


Figure 2 Neural Network with backpropagation approach tocontrol P3DX navigation from the start and exit of maze (source[157])

navigation from the start to the exit of maze as shown inFigure 2

ANN is believed to be very good at resolving nonlinearproblems that require mapping input and output relationwithout necessarily having knowledge of the system and itsenvironment Based on this strength Akkaya Aydogdu andCanan [167] proposed global positioning system (GPS) anddead reckoning (DR) sensor fusion approach by adoptingANN nonlinear autoregressive with external input (NARX)model to control mobile robot to avoid hitting obstacles Themobile robot kinematics was modelled using ANN Basedon the performance of the approach through simulation andexperiment the approach was presented by the authors asan alternative to conventional navigation methods that useKalman filter

To achieve intelligent mobile robot control in unknownenvironment Panigrahi et al [168] proposed a redial basisfunction (RBF)NN approach formobile robot path planningFarooq et al [169] also contributed by designing an intel-ligent autonomous vehicle capable of navigating noisy andunknown environment without collision using ANN Multi-layer feed forward NN controllers with back error propa-gation was adopted for robot safe path planning to reachgoal During the evaluation of the approach it was identifiedthat the efficiency of the neural controller deteriorates asthe number of layers increase due to the error term that isdetermined using approximated function

In another development Medina-Satiago et al [170]considered path planning as a problem of pattern classi-fication They developed a path planning algorithm usingmultilayer perceptron (MLP) and back propagation (BP) ofANN which was experimented on a differential drive mobilerobot Alternatively Syed et al [171] contributed by proposingguided autowave pulse coupled neural network (GAPCNN)which is an improvement of pulse coupled neural network(PCNN) by Qu et al [172] for mobile robot path planningand obstacle avoidance Dynamic thresholding techniquewas applied in their method Results from simulation andexperiments described the approach to be time efficient and

good for static and dynamic environments compared tomodified PCNN

In [173] a data-driven end-to-end motion planningmethod based on CNN was introduced to control a dif-ferential drive The method aimed at providing the abilityto learn navigation strategies Simulation and real experi-ment indicated that the model can be trained in simula-tion and transferred to a robotic platform to perform therequired navigation task in the real environment Localminima and oscillation problems of the method need to beaddressed

43 Genetic Algorithm Path PlanningMethods Another pop-ular method used by researchers over the years for mobilerobot path planning is GA GA is a metaheuristic inspired bythe process of natural selection that belongs to the larger classof evolutionary algorithms (EA) GA are normally consideredwhen there is the need to generate high-quality solutions tohelp optimization and search problems by using bio-inspiredoperators including mutation crossover and selection [174]GA is a category of search algorithms and optimizationmethods thatmake use ofDarwinrsquos evolutionary theory of thesurvival of the fittest [175] Research on the application of GAto robotic path planning and obstacle avoidance was done inthe past [38 176ndash184]

Tuncer and Yildirim [178] considered GA to present anew mutation operator for mobile robot path planning indynamic environment According to the authors comparingthe results to other methods showed better performanceof their technique in terms of path optimality With themotivation to provide optimal and collision freemobile robotnavigation a combinatorial optimization technique based onGA approach was used in [184] to propose bacterial memeticalgorithm to make a mobile robot navigate to a goal whileavoiding obstacles at a minimized path length

44 Memetic Algorithm Path Planning Method Memeticalgorithm is a category of evolutionary [185] method whichcombines evolutionary and local search methods Memeticalgorithm has been used to attempt solving optimizationproblems in [186] There is remarkable research work usingmemetic algorithm in mobile robot path planning [187 188]

One of suchworks is byBotzheimToda andKubota [189]They adopted the bacterial memetic algorithm in [190] byconsidering bacterial mutation and gene transfer operationas the operators The approach was applied to mobile robotnavigation and obstacle avoidance in static environment (seeFigure 3)

45 Particle Swarm Optimization Path Planning MethodPSO technique emerged from Kennedy and Eberhart [191]The technique was inspired by the social behavior of birdsand fish in groups by considering the best positions of eachbird or fish in search of food using fitness function parameterPSO is simple to implement and requires less computing timeand performs well in various optimization problems [192ndash195] PSO has been adopted by a considerable number of


Figure 3 Controlling mobile robot navigation using bacteriamemetic algorithm (source Botzheim Toda and Kubota [189])

researchers in mobile robot path planning task [192 196ndash208] The mathematical equations describing PSO [196] areshown in the following

119881119894 (119896 + 1) = 119908 lowast 119881119894 (119896) + 1198881 lowast 1199031 (119884119875119887119890119904119905 minus 119910119894) + 1198882lowast 1199032 (119884119875119887119890119904119905 minus 119910119894)

(4)

119910119894 (119896 + 1) = 119910119894 (119896) + 119881119894 (119896 + 1) (5)

where 119894 is particle number 119881119894 is velocity of the particle i 119910119894 isposition of the particle i 119896 is iteration counter 119908 is inertialweight (decreasing function) 1198881 and 1198882 are accelerationcoefficients known as cognitive and social parameters 1199031and 1199032 are random values in [0 1] 119884119875119887119890119904119905 is best position ofparticle and 119884119866119887119890119904119905 is global best position of particle

Recently Rath and Deepak [196] used PSO technique todevelop amotion planner for stationary andmovable objectsThe developed algorithmwas implemented in simulation andit is yet to be implemented in robotic platform to confirm theresults achieved in simulation

PSO approach applied in [209 210] was adopted from[205 206 210 211] for motion planning and obstacle avoid-ance in dynamic environment The approach was imple-mented in simulation environment

46 Artificial Bee Colony Path Planning Method ABC algo-rithm was also considered by few researchers [212] Basedon swarm intelligence and chaotic dynamic of learning Linand Huang [119] presented ABC algorithm for path planningfor mobile robots Though the paper stated the method wasmore efficient than GA and PSO it was not tested in realrobot platform to ascertain that fact Optimal path planningmethod based onABCwas also introduced in [116] to providecollision free navigation of mobile robot with the aim ofachieving optimal path Simulation was used to test thealgorithmwhichwas described to be successful However themethod was not demonstrated in experiment with real robot

Abbass and Ali [122] on their part presented what theydescribed as directed artificial bee colony algorithm (DABC)based on ABC algorithm for autonomous mobile robotpath planning The DABC algorithm was used to directthe bees to the direction of the best bee to help obtain

Figure 4 Best path results using DABC (source [122])

Figure 5 Best path results using ABC (source [122])

optimal path while avoiding obstacles Simulation resultscompared to conventional ABC algorithm was described tohave performed better as demonstrated in Figures 4 and 5

Recently bees algorithm was used in [213] to presenta real-time path planning method in an indoor dynamicenvironment The bees algorithm was used to generate thepath in static environment which is later updated onlineusing modified bees algorithm to enable preventing hittingobstacles in dynamic environment Neighborhood shrinkingwas used to optimize the performance of the algorithmSimulation and experiment using AmigoBot were used toevaluate the method and it was described to have performedwell with optimal path in a real time

47 Simulated Annealing Algorithm Path Planning MethodSA is a probabilistic technique used for estimating the globaloptimumof a function It was used in past research formobilerobot obstacle avoidance task [214ndash216] SA algorithm wasused in [217] to propose an algorithm to enablemobile robots


to reach a goal through dynamic environment composed ofobstacles with optimal path This method was implementedsuccessfully in simulation

48 Ant Colony Optimization Algorithm Path PlanningMethod ACO is a probabilistic method used for findingsolution for computational problems including path plan-ning ACO technique has been considered by researchersfor mobile robot path planning over the years [29 218ndash222]Using ACO Guan-Zheng et al [220] demonstrated throughalgorithm and simulation a path planningmethod for mobilerobots Results were compared to GA method and it wasindicated that the method was effective and can be used inthe real-time path planning of mobile robots

Vien et al [221] used Ant-Q reinforcement learningbased on Ant Colony approach algorithm as a technique toaddress mobile robot path planning and obstacle avoidanceproblem Results from simulationwere comparedwith resultsfrom other heuristic based on GA Comparatively Ant-Qreinforcement learning approach was described to be betterin terms of convergence rate

Considering the complex obstacle avoidance probleminvolving multiple mobile robots Ioannidis et al [222] pro-posed a path planner usingCellular Automata (CA) andACOtechniques to provide collision free navigation for multiplerobots operating in the same environment while keepingtheir formation immutable Implementation was done in asimulated environment consisting of cooperative team ofrobots and an obstacle

Real robot platform experimental implementation maybe required to prove the efficiency of these approaches toconfirm the simulation results

49 Human Navigation Dynamics Path Planning MethodsHuman navigation dynamics (HND) research was also givenattention by researchers [68 223ndash230] However very fewconsiderations were given to the application of HND strate-gies to path planning and obstacle avoidance of mobilerobots HND strategy termed as all-or-nothing strategy wasused by Frank et al [231] to propose aminimalistic navigationmodel based on complex number calculus to solve mobilerobotsrsquo navigation and obstacle avoidance problem Thisapproach did not however demonstrate any implementationthrough experiments Much work may be required in thisarea to explore and experiment the use of human navigationstrategies to robotics motion planning and obstacle avoid-ance

5 Strengths and Challenges ofNature-Inspired Path Planning Methods

51 Strengths of Nature-Inspired Path Planning MethodsNature-inspired path planning methods possess the abilityto imitate the behavior of some living things that havenatural intelligence Methods like ANN and FL are good atproviding learning and generalization [232] FL for instancepossess the ability to make inferences in uncertain sce-narios When combined with FL method ANN can tunethe rule base of FL to produce better results The learning

component of these methods aids in effective performancein unknown and dynamic environment of the autonomousvehicle GA and other nature-inspired methods have goodoptimization ability and are good at solving problems thatare difficult dealing with using conventional approachesACO method and memetic algorithms are noted for theirfast convergence characteristics and good at obtainingoptimal result [233ndash235] Moreover nature-inspired meth-ods can combine well with other optimization algorithms[236 237] to provide efficient path planning in the realenvironment

52 Challenges of Nature-Inspired Path Planning MethodsNotwithstanding the strengths discussed above there aresome weaknesses of nature-inspired path planning methodssome of which include trapping in local minima slowconvergence speed premature convergence high computingpower requirement oscillation difficulty in choosing initialpositions and the requirement of large data set of theenvironment which is difficult to obtain

Despite the strength of ANN stated in the previoussection they require large set of data of the environmentduring training to achieve the best result which is difficultto obtain especially with supervised learning [232] The useof backpropagation technique to provide efficient algorithmalso have its challenges BP method easily converges to localminima if the situation actionmapping is not convex with theparameters [238] The algorithm stops at local minima if itis above its global minimum Moreover it is also describedto have very slow convergence speed which may result insome collisions before the robot gets to its defined goal[232] FL systems can provide knowledge-based heuristicsituation action mapping however their rule bases are hardto build in unstructured environment [232] This makes itdifficult using FL method to address path planning problemsin unstructured environments without combining it withother methods Despite the strength of good optimizationcapacity of GA it is difficult for GA to scale well withcomplex scenarios and it is characterized with convenienceat local minima and oscillation problems [239 240] Alsodue to its complex principle it may be difficult to deal withdynamic data sets to achieve good results [233] Though PSOis described to be simple with less computing time require-ment and effective in implementing with varied optimizationproblems with good results [192ndash195] it is difficult to dealwith trapping into local minima problems under complexmap [194 240] Although ACO is noted for fast convergencewith optimal results [233 234] it requires a lot of computingtime and it is difficult to determine the parameters whichaffects obtaining quick convergence [233 240] Comparedto conventional algorithms memetic algorithm is describedto possess the ability to produce faster convergence andgood result however it can result in premature convergence[235] ABC is simple with fewer control parameters with lesscomputational time but low convergence is a drawback ofABC algorithm [241] It is believed that SA performs wellin approximating the global optimum but the algorithm isslow and it is difficult choosing the initial position for SA[242]


(a) (b)

Figure 6 Mobile robot obstacle avoidance using DRNFS with short memory (a) compared to conventional fuzzy rule-based (b) navigationsystem from a given start (S) and goal (G) positions (source [261])

6 Hybrid Methods

To take advantage of the strengths of some methods whilereducing the effects of their drawbacks some researchershave combined two or more methods in their investigationsto provide efficient hybrid path planning method to controlautonomous ground vehicles Some of these approachesinclude neuro-fuzzy wall-following-based fuzzy logic fuzzylogic combined with Kalman Filter APF combined with GAand APF combined with PSO Some of the hybrid approachesare discussed in the next subsections

61 Neuro-Fuzzy Path Planning Method Popular amongthe hybrid approaches is neuro-fuzzy also termed as fuzzyneural network (FNN) It is a combination of ANN and FLThis approach considers the human-like reasoning style offuzzy systems using fuzzy sets and linguistic model that arecomposed of if-then fuzzy rules as well as ANN [243] Theuse of neuro-fuzzy system approach in obstacle avoidancefor mobile robotsrsquo research has been considered by manyresearchers [101 244ndash260]

A weightless neural network (WNN) approach usingembedded interval type-2 neuro-fuzzy controller with rangesensor to detect and avoid obstacles was presented in[256 257] This approach worked but its performance wasdescribed to be interfered by noise Unified strategies of pathplanning using type-1 fuzzy neural network (T1FNN) wasproposed in [16] to achieve obstacle avoidance It is howeveridentified in [101] that the use of (T1FNN) have challengesincluding the unsatisfactory control performance and thedifficulty of reducing the effect of uncertainties in achievingtask and the oscillation behavior that is characterized bythe approach during obstacle avoidance Kim and Chwa [14]took advantage of this limitation and modified the TIFNNto propose an obstacle avoidance method using intervaltype-2 fuzzy neural network (IT2FNN) Evaluation of thisstrategy was carried out using simulation and experimentand compared with the T1FNN approach The IT2FNN wasdescribed to have produced better results

AbuBaker [258] presented a neuro-fuzzy strategy termedas neuro-fuzzy optimization controller (NFOC) to controlmobile robot to avoid colliding with obstacles while movingto goal NFOC approach was evaluated in simulation and wascompared with fuzzy logic controller (FLC40) and FLC625and the performance was described to be better in terms ofresponse time

Chen and Richardson [261] on the other hand tried tolook at path planning and obstacle avoidance of mobile robotin relation to the human driver point of view They adopteda new fuzzy strategy to propose a dynamic recurrent neuro-fuzzy system (DRNFS) with short memory Their methodwas simulated using three ultrasonic sensors to collect datafrom the environment 18 obstacle avoidance trajectories and186 training data pairs to train the DRNFS and comparedto conventional fuzzy rule-based navigation system and theapproach was described to be better in performance Figure 6demonstrates the performance of the DRNFS compared tothe conventional fuzzy rule-based navigation system

Another demonstration of neuro-fuzzy approach waspresented in [259] with backpropagation algorithm to drivemobile robot to avoid obstacles in a challenging environmentSimulation and experimental results showed partial solutionof reduction of inaccuracies in steering angle and reachinggoal with optimal path A proposal using FL control com-bined with artificial neural network was presented by Jeffriland Sariff [260] to control the navigation of a mobile robot toavoid obstacles

To capitalize on the strengths of neural network and FLAlgabri Mathkour and Ramdane [262] proposed adaptiveneuro-fuzzy inference system (ANFIS) approach for mobilerobot navigation and obstacle avoidance Simulation wascarried out using Khepera simulator (KiKs) in MATLAB asshown in Figure 7 Practical implementation was also doneusing Khepera robot in an environment scattered with fewobstacles

A combination of FL and spiking neural networks (SNN)approach was proposed in [263] and simulation results was


Figure 7 Khepera robot navigation in scattered environment withobstacles using ANFIS approach (source [262])

compared to that of FL The new approach was described tohave performed better Alternatively ANFIS controller wasdeveloped using neural network approach to controlmultiplemobile robot to reach their target while avoiding obstacles ina dynamic indoor environment [18 264ndash266] The authorsdeveloped ROBNAV software to handle the control of mobilerobot navigation using the ANFIS controller developed

Pothal and Parhi [267] developedAdaptiveNeuron FuzzyInference System (ANFIS) controller for mobile robot pathplanning Implementationwas done using single robot aswellasmultiple robotsThe average percentage error of time takenfor a single robot to reach its target comparing simulation andexperiment results was 1047

To deal with the effect of noise generation during infor-mation collection using sensors a Hybrid Intelligent System(HIS) was proposed by Alves and Lopes [268] They adoptedFL and ANN to control the navigation of the robot While inneuro-fuzzy system training needs to be done from scratchthis method deviated a little bit where only the fuzzy moduleis calibrated and trained Simulation results showed that themethod was successful

Realizing the challenge of training and updating param-eters of ANFIS in path planning which usually leads tolocal minimum problem teaching-learning-based optimiza-tion (TLBO) was combined with ANFIS to present a pathplanning method in [269] The aim was to exploit thebenefits of the two methods to obtain the shortest pathand time to goal in strange environment The TLBO wasused to train the ANFIS structure parameters without seek-ing to adjust parameters PSO invasive weed optimiza-tion (IWO) and biogeography-based optimization (BBO)algorithms were also used to train the ANFIS parametersto aid comparison with the proposed method Simula-tion results indicated that TLBO-based ANFIS emergedwith better path length and time Improvement may berequired to consider other path planning tasks It may alsorequire comparison with other path planning methods todetermine its efficiency Real experiment should be consid-ered to prove the effectiveness of the method in the realenvironment

62 Other Hybrid Methods That Include FL Wall-following-based FL approach has also been considered by researchersOne of the pioneers among them is Braunstingl Anz-JandEzkera [270]They used fuzzy logic rules to implement awall-following-based obstacle controller Wall-following controllaw based on interval type-2 fuzzy logic was also proposed in[131 271 272] for path planning and obstacle avoidance whiletrying to improve noise resistance ability Recently Al-Mutiband Adessemed [273] tried to address the oscillation anddata uncertainties problems and proposed a fuzzy logic wall-following technique based on type-2 fuzzy sets for obstacleavoidance for indoor mobile robots The fuzzy controllerproposed relied on the distance and orientation of the robotto the object as inputs to generate the needed wheel velocitiestomove the robot smoothly to its target while switching to theobstacle avoidance algorithm designed to avoid obstacles

Despite the problem of sensor data uncertainties Hankand Haddad [274] took advantage of the strengths of FL andcombined trajectory-tracking and reactive-navigation modestrategy with fuzzy controller used in [275] for mobile robotpath planning Simulation and experiments were performedto test the approach which was described to have producedgood results

In [276] an approach called dynamic self-generated fuzzyQ-learning (DSGFQL) and enhanced dynamic self-generatedfuzzy Q-learning (EDSGFQL) were proposed for obstacleavoidance task The approach was compared to fuzzy Q-learning (FQL) [277] dynamic fuzzy Q-learning (DFQL)of Er and Deng [135] and clustering and Q-value basedGA learning scheme for fuzzy system design (CQGAF) ofJuang [278] and it was proven to perform better throughsimulation Considering a learning technique in a differentdirection a method termed as reinforcement ant optimizedfuzzy controller (RAOFC) was designed by Juang and Hsu[279] for wheeled mobile robot wall-following under rein-forcement learning environments This approach used anonline aligned interval type-2 fuzzy clustering (AIT2FC)method to generate fuzzy rules for the control automaticallyThismakes it possible not to initialize the fuzzy rules Q-valueaided ant colony optimization (QACO) technique in solvingcomputational problemswas employed in designing the fuzzyrules The approach was implemented considering the wallas the only obstacle to be avoided by the robot in an indoorenvironment

Recently a path planning approach for goal seeking inunstructured environment was proposed using least mean p-norm extreme learningmachine (LMP-ELM) andQ-learningby Yang et al [232] LMP-ELM and Q-learning are neuralnetwork learning algorithm To control errors that may affectthe performance of the robots least mean P-power (LMP)error criterion was introduced to sequentially update theoutput Simulation results indicated that the approach isgood for path planning and obstacle avoidance in unknownenvironment However no real experiment was conducted toevaluate the method

Considering the limitation of conventional APF methodwith the inability of mobile robots to pass between closedspace Park et al [280] combined fuzzy logic and APFapproaches as advanced fuzzy potential field method


1 procedure NAVIGATION(qo qf ON 120578 e M Np)2 ilarr997888 03 navigationlarr997888 True4 larr997888 BPF(qo qf ON 120578 e M Np) is an array5 while navigation do6 Perform verification of the environment7 if environment has changed then8 q119900 larr997888 (i)9 larr997888 BPF(qo qf ON 120578 e M Np)10 ilarr997888 011 end if12 ilarr997888 i+ 113 Perform trajectory planning to navigate from (i-1) to (i)14 if i gt= length of then15 navigationlarr997888 False16 Display Goal has been achieved17 end if18 end while19 end procedure

Algorithm 1 BPF algorithm

(AFPFM) to address this limitation The position andorientation of the robot were considered in controlling therobot The approach was evaluated using simulation

To address mobile robot control in unknown environ-ment Lee et al [281] proposed sonar behavior-based fuzzycontroller (BFC) to control mobile robot wall-following Thesub-fuzzy controllers with nine fuzzy rules were used for thecontrol The Pioneer 3-DX mobile robot was used to evaluatethe proposed method Although the performance was goodin the environment with regular shaped obstacles collisionswere recorded in the environment with irregular obstaclesMoreover the inability to ensure consistent distance betweenthe robot and the wall is a challenge

A detection algorithm for vision systems that combinesfuzzy image processing algorithm bacterial algorithm GAand Alowast was presented in [282] to address path planningproblem The fuzzy image processing algorithm was used todetect the edges of the images of the robotrsquos environmentThe bacterial algorithm was used to optimize the output ofthe processed image The optimal path and trajectory weredetermined using GA and Alowast algorithms The results ofthe method indicate good performance in detecting edgesand reducing the noise in the images This method requiresoptimization to control performance in lighting environmentto deal with reflection from images

63 Other Hybrid Path Planning Methods There are manyother hybrid approaches used for path planning that did notinclude FL Some of these include APF combined with SA[216] PSO combined with gravitational search (GS) algo-rithm [283] VFH algorithm combined with Kalman Filter[113] integrating game theory and geometry [284] com-bination of differential global positioning system (DGPS)APF FL and Alowast path planning algorithm [41] Alowast searchand discrete optimization methods [92] a combination ofdifferential equation and SMC [243] virtual obstacle concept

was combined with APF [106] and recently the use of LIDARsensor accompanied with curve fitting Few of such methodsare discussed in this section

Recently Marin-Plaza et al [285] proposed and imple-mented a path planning method based on Ackermannmodelusing time elastic bands Dijkstra method was employed toobtain the optimal path for the navigation A good result wasachieved in a real experiment conducted using iCab for thevalidation of the method

Considering the strengths and weaknesses of APF in goalseeking and obstacle avoidance Montiel et al [6] combinedAPF and bacterial evolutionary algorithm (BEA) methods topresent Bacterial Potential Field (BPF) to provide safe andoptimal mobile robot navigation in complex environmentThe purpose of the strategy was to eliminate the trappingin local minima drawbacks of the traditional APF methodSimulation results was described to have showed betterresults compared to genetic potential field (GPF) and pseudo-bacterial potential field (PBPF)methods Unknown obstacleswere detected and avoided during the simulation (See Fig-ure 8) The BPF algorithm in [6] is given in Algorithm 1

Experiment demonstrated that BPF approach comparedto other APF methods used a lower computational time of afactor of 159 to find the optimal path

APF method was optimized using PSO algorithm in [36]to develop a control system to control the local minimaproblem of APF Implementation of this approach was doneusing simulation as demonstrated in Figure 9 Real robotplatform implementation may be required to confirm theeffectiveness of the technique demonstrated in simulation

A visual-based approach using a camera and a rangesensorwas presented in [286] by combiningAPF redundancyand visual servoing to deal with path planning and obstacleavoidance problem of mobile robots This approach wasimproved upon in [287] by including visual path following


Figure 8 Unknown obstacle detected during navigation using BPFapproach (source [6])

Figure 9 Controlling the local minima problem of APF in pathplanning and obstacle avoidance by optimizing APF using PSOalgorithm (source [36])

obstacle bypassing and collision avoidance with decelerationapproach [288 289] The method was designed to enablerobots progress along a path while avoiding obstacles andmaintaining closeness to the 3D path Experiment was donein outdoor environment using a cycab robot with a 70 fieldof view Also based on extended Kalman filters a classicalmotion stereo vision approach was demonstrated in [290] forvisual obstacle detection which could not be detected by laserrange finders Visual SLAM algorithm was also proposedin [291] to build a map of the environment in real timewith the help of Field Programmable Gate Arrays (FPGA) toprovide autonomous navigation ofmobile robots in unknownenvironment Since navigation involves obstacle avoidanceAPF method was presented for the obstacle avoidance Thealgorithm was tested in indoor environment composed ofstatic and dynamic obstacles using a differential mobile robotwith cameras Different from the normal Kalman filterscomputed depth estimates derived from traditional stereomethodwere used to initialize the filters instead of using sim-ple heuristics to choose the initial estimates Target-driven

visual navigation method was presented in [292] to addressthe impractical application problem of deep reinforcementlearning in real-world situations The paper attributed thisproblem to lack of generalization capacity to new goals anddata inefficiency when using deep reinforcement learningActor-critic andAI2-THORmodelswere proposed to addressthe generalization and data efficiency problems respectivelySCITOS robot was used to validate the method Thoughresults showed good performance a test in environmentcomposed of obstacles may be required to determine itsefficiency in the real-world settings

A biological navigation algorithm known as equiangularnavigation guidance (ENG) law approach accompanied withthe use of local obstacle avoidance techniques using sensorswas employed in [13] to plan the motion of a robot toavoid obstacles in complex environmentwithmany obstaclesThe choice of the shortest path to goal was however notconsidered

Savage et al [293] also presented a strategy combiningGA with recurrent neural network (RNN) to address pathplanning problem GA was used to find the obstacle avoid-ance behavior and then implemented in RNN to control therobot To address path planning problem in unstructuredenvironment with relation to unmanned ground vehiclesMichel et al [94] presented a learning algorithm approach bycombining reinforcement learning (RL) computer graphicsand computer vision Supervised learning was first used toapproximate the depths from a single monocular imageThe proposed algorithm then learns monocular vision cuesto estimate the relative depths of the obstacles after whichreinforcement learning was used to render the appropriatesynthetic scenes to determine the steering direction of therobot Instead of using the real images and laser scan datasynthetic images were used Results showed that combiningthe synthetic and the real data performs better than trainingthe algorithm on either synthetic or real data only

To provide effective path planning and obstacle avoidancefor a mobile robot responsible for transporting componentsfrom a warehouse to production lines Zaki Arafa and Amer[294] proposed an ultrasonic ranger slave microcontrollersystem using hybrid navigation system approach that com-bines perception and dead reckoning based navigation Asmany as 24 ultrasonic sensors were used in this approach

In [295] the authors demonstrated the effectiveness ofGA and PRM path planning methods in simple and complexmaps Simulation results indicated that both methods wereable to find the path from the initial state to the goalHowever the path produced by GA was smoother than thatof PRM Again comparing GA and PRM PRM performedpath computation in lesser timewhichmakes itmore effectivein reactive path planning applications GA and PRM couldbe combined to exploit the benefits of the two methodsto provide smooth and less time path computation Realexperiment is required to evaluate the effectiveness of themethods in a real environment

Hassani et al [296] aimed at obtaining safe path incomplex environment for mobile robot and proposed analgorithm combining free segments and turning points algo-rithms To provide stability during navigation sliding mode


control was adopted Simulation in Khepera IV platformwas used to evaluate the method using maps of knownenvironment composed of seven static obstacles Resultsindicated that safe and optimal path was generated fromthe initial position to the target This method needs to becompared to other methods to determine its effectivenessMoreover a test is required in a more complex environmentand in an environment with dynamic obstacles Issues ofreplanning should be considered when random obstacles areencountered during navigation

Path planning approach to optimize the task of mobilerobot path planning using differential evolution (DE) andBSpline was given in [297] The path planning task was doneusing DE which is an improved form of GA To ensureeasy navigation path the BSpline was used to smoothenthe path Aria P3-DX mobile robot was used to test themethod in a real experiment Compared to PSO method theresults showed better optimality for the proposed methodAnother hybrid method that includes BSpline is presentedin [298] The authors combined a global path planner withBSpline to achieve free and time-optimal motion of mobilerobots in complex environmentsThe global path plannerwasused to obtain the waypoints in the environment while theBSpline was used as the local planner to address the optimalcontrol problem Simulation results showed the efficiencyof the method The KUKA youBot robot was used for realexperiment to validate the approach but was carried out insimple environment A test in a complex environment maybe required to determine the efficiency of the method

Recently nonlinear kinodynamic constraints in pathplanning was considered in [299] with the aim of achievingnear-optimalmotion planning of nonlinear dynamic vehiclesin clustered environment NN and RRT were combined topropose NoD-RRT method RRT was modified to performreconstruction to address the kinodynamic constraint prob-lem The prediction of the cost function was done using NNThe proposed method was evaluated in simulation and realexperiment using Pioneer 3-DX robot with sonar sensorsto detect obstacles Compared to RRT and RRTlowast it wasdescribed to have performed better

7 Strengths and Challenges of Hybrid PathPlanning Methods

Because hybrid methods are combination of multiple meth-ods they possess comparatively higher merits by takingadvantage of the strengths of the integrated methods whileminimizing the drawbacks of these methods For instanceintegration of appropriate methods can help improve noiseresistance ability and deal better with oscillation and datauncertainties and controlling local minima problem associ-ated with APF [36 131 171 272 273]

Though hybrid methods seem to possess the strengths ofthemethods integratedwhile reducing their drawbacks thereare still challenges using them Compatibility of some of thesemethods is questionable The integration of incompatiblemethods would produce worse results than using a singlemethod Other weaknesses not noted with the individualmethods combined may emerge when the methods are

integrated and implemented New weaknesses that emergebecause of the combination may also lead to unsatisfactorycontrol performance and difficulty of reducing the effects ofuncertainty and oscillations Noise from sensors and camerasin addition to other hardware constraints including thelimitation of motor speed imbalance mass of the robotunequal wheel diameter encoder sampling errors misalign-ment of wheels disproportionate power supply to themotorsuneven or slippery floors and many other systematic andnonsystematic errors affects the practical performance ofthese approaches

8 Conclusion and the Way Forward

In conclusion achieving intelligent autonomous groundvehicles is the main goal in mobile robotics Some out-standing contributions have been made over the yearsin the area to address the path planning and obstacleavoidance problems However path planning and obstacleavoidance problem still affects the achievement of intelligentautonomous ground vehicles [77 145] In this paper weanalyzed varied approaches from some outstanding publica-tions in addressing mobile robot path planning and obstacleavoidance problems The strengths and challenges of theseapproaches were identified and discussed Notable amongthem is the noise generated from the cameras sensors andthe constraints of themobile robotswhich affect the reliabilityand efficiency of the approaches to perform well in realimplementations Localminima oscillation andunnecessarystopping of the robot to update its environmental mapsslow convergence speed difficulty in navigating in clutteredenvironment without collision and difficulty reaching targetwith safe optimum path were among the challenges identi-fied It was also identified that most of the approaches wereevaluated only in simulation environment The challenge ishow successful or efficient these approaches would be whenimplemented in the real environment where real conditionsexist [266] A summary of the strengths and weaknesses ofthe approaches considered are shown in Tables 1 2 and 3

The following general suggestions are given in thispaper when proposing new methods for path planning forautonomous vehicles Critical consideration should be givento the kinematic dynamics of the vehicles The systematicand nonsystematic errors that may affect navigation shouldbe looked at Also the limitation of the environmental datacollection devices like sensors and cameras needs to beconsidered since their limitation affects the information to beprocessed to control the robots by the proposed algorithmThere is therefore the need to provide a method that can con-trol noise generated from these devices to aid best estimationof data to control and achieve safe optimal path planningTo enable the autonomous vehicle to take quick decision toavoid collision into obstacles the computation complexity ofalgorithms should be looked at to reduce the execution timeMoreover the strengths and limitations of an approach or amethod should be looked at before considering its implemen-tation Possibly hybrid approaches that possess the ability totake advantage of the benefits and reduce the limitations ofeach of the methods involved can be considered to obtain


Table 1 Summary of Strengths and Challenges of Nature-inspired computation based mobile robot path planning and obstacle avoidancemethods

Approach Major Imple-mentation Strengths Challenges

Neural Network Simulation andreal Experiment

(i) Good at providing learning and generalization[232](ii) Can help tune the rule base of fuzzy logic [232]

(i) Efficiency of neural controllers deteriorates as thenumber of layers increase [232](ii) Difficult to acquire its required large dataset ofthe environment during training to achieve bestresults(iii) Using BP easily results in local minima problems[238](iv) It has a slow convergence speed [232]

GeneticAlgorithm Simulation

(i) Good at solving problems that are difficult todeal with using conventional algorithms and itcan combine well with other algorithms(ii) It has good optimization ability

(i) Difficult to scale well with complex situations(ii) Can lead to convergence at local minima andoscillations [239 240](iii) Difficult to work on dynamic data sets anddifficult to achieve results due to its complexprinciple [233]

Ant Colony Simulation (i) Good at obtaining optimal result [233](ii) Fast convergence [234]

(i) Difficult to determine its parameters which affectsobtaining quick convergence [240](ii) Require a lot of computing time [233]

Particle SwarmOptimization Simulation

(i) Faster than fuzzy logic in terms of convergence[132](ii) Simple and does not require much computingtime hence effective to implement withoptimization problems [192ndash195](iii) Performs well on varied application problems[193]

(i) Difficult to deal with trapping into local minimaunder complex map [194 240](ii) Difficult to generalize its performance sinceundertaken experiments relied on objects in polygonform [240]

Fuzzy LogicSimulation andExperiments

with real robot

(i) Ability to make inferences in uncertainscenarios [129](ii) Ability to imitate the control logic of human[129](ii) It does well when combine with otheralgorithms

(i) Difficult to build rule base to deal withunstructured environment [145 232]

Artificial BeeColony Simulation

(i) It does not require much computational timeand it produces good results [241](ii) It has simple algorithm and easy to implementto solve optimization problems [236 241](iii) Can combine well with other optimizationalgorithms [236 237](iv) It uses fewer control parameters [236]

(i) It has low convergence performance [241]

Memetic andBacterialMemeticalgorithm

Simulation

(i) Compared to conventional algorithms itproduces faster convergence and good solutionwith the benefit from different search methods itblends [235]

(i) It can result in premature convergence [235]

Others(SimulatedAnnealing andHumanNavigationstrategy)

Simulation

(i) Simulated Annealing is good at approximatingthe global optimum [242](ii) Takes advantage of human navigation thatdoes not depend on explicit planning but occuronline [223]

(i) SA algorithm is slow [242](ii) Difficult in choosing the initial position for SA[242]

better techniques for path planning and obstacle avoidancetask for autonomous ground vehicles Again efforts shouldbe made to implement proposed algorithms in real robotplatform where real conditions are found Implementationshould be aimed at both indoor and outdoor environments

to meet real conditions required of intelligent autonomousground vehicles

To achieve safe and efficient path planning of autonomousvehicles we specifically recommend a hybrid approachthat combines visual-based method morphological dilation


Table 2 Summary of strengths and challenges of conventional based mobile robot path planning and obstacle avoidance methods


ArtificialPotential Field

Simulation andreal Experiment (i) Easy to implement by considering attraction to

goal and repulsion to obstacles

(i) Results in local minima problem causing therobot to be trapped at a position instead of the goal[36 40 41 105 106](ii) Inability of the robot to pass between closelyspaced obstacles and results in oscillations[107 108 110](iii) Difficulty to perform in environment withobstacles of different shapes [109](iv) Constraints of the hardware of the mobile robotsaffect the performance of APF methods [51]

Visual Basedand Reactiveapproaches

Simulation andreal Experiment

(i) Easy to implement by relying on theinformation from sensors and cameras to takedecision on the movement of the robot

(i) Noise from sensors and cameras due to distancefrom objects color temperature and reflectionaffects performance(ii) Hardware constraints of the robots affectperformance(iii) Computational expensive

Robot Motionand SlidingMode Strategy

Simulation(i) Is Fast with respect to response time(ii) Works well with uncertain systems and otherunconducive external factors [64]

(i) Performs poorly when the longitudinal velocity ofthe robot is fast(ii) Computational expensive(iii) Chattering problem leading to low controlaccuracy [114]

DynamicWindow Simulation (i) Easy to implement (i) Local minima problem

(ii) Difficult to manage the effects of mobile robotconstraints [75 102ndash104]

Others (KFSGBM Alowast CSSGS Curvaturevelocity methodBumper Eventapproach Wallfollowing)

Simulation andExperiments

with real robot(i) Easy to implement (i) Noise from sensors affects performance

(ii) Computational expensive

Table 3 Summary of strengths and challenges of hybrid mobile robot path planning and obstacle avoidance methods


Neuro FuzzySimulation andreal Experiment

Takes advantage of the strengthsof NN and Fuzzy logic whilereducing the drawbacks of bothmethods Example NN can helptune the rule base of fuzzy logicwhich is difficult using fuzzylogic alone [145 232]

Unsatisfactory controlperformance and difficulty ofreducing the effect of uncertaintyand oscillation of T1FNN [101]

Others (Kalman Filter and Fuzzylogic Visual Based and APF Wallfollowing and Fuzzy Logic Fuzzylogic and Q-Learning GA andNN APF and SA APF andVisual Servoing APF and AntColony Fuzzy and AlowastReinforcement and computergraphics and computer visionPSO and Gravitational searchetc)

Simulation andreal

Experiments

The drawbacks of each approachin the combination is reducedExample Improves noiseresistance ability deal better withoscillation and data uncertainties[131 271ndash273] and controllinglocal minima problem associatedwith APF [36]

Noise from sensor and camerasand the hardware constraintsincluding the limitation of motorspeed imbalance mass of therobot power supply to themotors and many others affectsthe practical performance ofthese approaches


(MD) VD neuro-fuzzy Alowast and path smoothing algorithmsThe visual-based method is to help in collecting and process-ing visual data from the environment to obtain the map ofthe environment for further processing To reduce uncertain-ties of data collection tools multiple cameras and distancesensors are required to collect the data simultaneously whichshould be taken through computations to obtain the bestoutput The MD is required to inflate the obstacles in theenvironment before generating the path to ensure safety ofthe vehicles and avoid trapping in narrow passages Theradius for obstacle inflation using MD should be based onthe size of the vehicle and other space requirements to takecare of uncertainties during navigation The VD algorithm isto help in constructing the roadmap to obtain feasible waypoints VD algorithm which finds perpendicular bisector ofequal distance among obstacle points would add to providingsafety of the vehicle and once a path exists it would find itThe neuro-fuzzymethodmay be required to provide learningability to deal with dynamic environment To obtain theshortest path Alowastmay be used and finally a path smootheningalgorithm to get a smooth navigable path The visual-basedmethod should also help to provide reactive path planningin dynamic environment While implementing the hybridmethod the kinematic dynamics of the vehicle should betaken into consideration This is a task we are currentlyinvestigating

This study is necessary in achieving safe and optimalnavigation of autonomous vehicles when the outlined rec-ommendations are applied in developing path planningalgorithms

Conflicts of Interest

The authors declare that there are no conflicts of interestregarding the publication of this paper

References

[1] S G Tzafestas ldquoMobile Robot PathMotion andTask Planningrdquoin Introduction of Mobile Robot Control Elsevier Inc pp 429ndash478 Elsevier Inc 2014

[2] Y C Kim S B Cho and S R Oh ldquoMap-building of a realmobile robot with GA-fuzzy controllerrdquo vol 4 pp 696ndash7032002

[3] S M Homayouni T Sai Hong and N Ismail ldquoDevelopment ofgenetic fuzzy logic controllers for complex production systemsrdquoComputers amp Industrial Engineering vol 57 no 4 pp 1247ndash12572009

[4] O R Motlagh T S Hong and N Ismail ldquoDevelopment of anew minimum avoidance system for a behavior-based mobilerobotrdquo Fuzzy Sets and Systems vol 160 no 13 pp 1929ndash19462009

[5] A S Matveev M C Hoy and A V Savkin ldquoA globallyconverging algorithm for reactive robot navigation amongmoving and deforming obstaclesrdquoAutomatica vol 54 pp 292ndash304 2015

[6] OMontiel U Orozco-Rosas and R Sepulveda ldquoPath planningfor mobile robots using Bacterial Potential Field for avoidingstatic and dynamic obstaclesrdquo Expert Systems with Applicationsvol 42 no 12 pp 5177ndash5191 2015

[7] OMotlagh S H Tang N Ismail and A R Ramli ldquoA review onpositioning techniques and technologies A novel AI approachrdquoJournal of Applied Sciences vol 9 no 9 pp 1601ndash1614 2009

[8] L Yiqing Y Xigang and L Yongjian ldquoAn improved PSOalgorithm for solving non-convex NLPMINLP problems withequality constraintsrdquo Computers amp Chemical Engineering vol31 no 3 pp 153ndash162 2007

[9] B B V L Deepak D R Parhi and B M V A Raju ldquoAdvanceParticle Swarm Optimization-Based Navigational ControllerForMobile RobotrdquoArabian Journal for Science and Engineeringvol 39 no 8 pp 6477ndash6487 2014

[10] S S Parate and J L Minaise ldquoDevelopment of an obstacleavoidance algorithm and path planning algorithm for anautonomous mobile robotrdquo nternational Engineering ResearchJournal (IERJ) vol 2 pp 94ndash99 2015

[11] SM LaVallePlanningAlgorithms CambridgeUniversity Press2006

[12] W H Huang B R Fajen J R Fink and W H Warren ldquoVisualnavigation and obstacle avoidance using a steering potentialfunctionrdquo Robotics and Autonomous Systems vol 54 no 4 pp288ndash299 2006

[13] H Teimoori and A V Savkin ldquoA biologically inspired methodfor robot navigation in a cluttered environmentrdquo Robotica vol28 no 5 pp 637ndash648 2010

[14] C-J Kim and D Chwa ldquoObstacle avoidance method forwheeled mobile robots using interval type-2 fuzzy neuralnetworkrdquo IEEE Transactions on Fuzzy Systems vol 23 no 3 pp677ndash687 2015

[15] D-H Kim and J-H Kim ldquoA real-time limit-cycle navigationmethod for fast mobile robots and its application to robotsoccerrdquo Robotics and Autonomous Systems vol 42 no 1 pp 17ndash30 2003

[16] C J Kim M-S Park A V Topalov D Chwa and S K HongldquoUnifying strategies of obstacle avoidance and shooting forsoccer robot systemsrdquo in Proceedings of the 2007 InternationalConference on Control Automation and Systems pp 207ndash211Seoul South Korea October 2007

[17] C G Rusu and I T Birou ldquoObstacle Avoidance Fuzzy Systemfor Mobile Robot with IRrdquo in Proceedings of the Sensors vol no10 pp 25ndash29 Suceava Romannia 2010

[18] S K Pradhan D R Parhi and A K Panda ldquoFuzzy logictechniques for navigation of several mobile robotsrdquoApplied SoftComputing vol 9 no 1 pp 290ndash304 2009

[19] H-J Yeo and M-H Sung ldquoFuzzy Control for the ObstacleAvoidance of Remote Control Mobile Robotrdquo Journal of theInstitute of Electronics Engineers of Korea SC vol 48 no 1 pp47ndash54 2011

[20] M Wang J Luo and U Walter ldquoA non-linear model predictivecontroller with obstacle avoidance for a space robotrdquo Advancesin Space Research vol 57 no 8 pp 1737ndash1746 2016

[21] Y Chen and J Sun ldquoDistributed optimal control formulti-agentsystems with obstacle avoidancerdquo Neurocomputing vol 173 pp2014ndash2021 2016

[22] T Jin ldquoObstacle Avoidance of Mobile Robot Based on BehaviorHierarchy by Fuzzy Logicrdquo International Journal of Fuzzy Logicand Intelligent Systems vol 12 no 3 pp 245ndash249 2012

[23] C Giannelli D Mugnaini and A Sestini ldquoPath planning withobstacle avoidance by G1 PH quintic splinesrdquo Computer-AidedDesign vol 75-76 pp 47ndash60 2016

[24] A S Matveev A V Savkin M Hoy and C Wang ldquoSafe RobotNavigation Among Moving and Steady Obstaclesrdquo Safe RobotNavigationAmongMoving and SteadyObstacles pp 1ndash344 2015


[25] S Jin and B Choi ldquoFuzzy Logic System Based ObstacleAvoidance for a Mobile Robotrdquo in Control and Automationand Energy System Engineering vol 256 of Communicationsin Computer and Information Science pp 1ndash6 Springer BerlinHeidelberg Berlin Heidelberg 2011

[26] D Xiaodong G A O Hongxia L I U Xiangdong and Z Xue-dong ldquoAdaptive particle swarm optimization algorithm basedon population entropyrdquo Journal of Electrical and ComputerEngineering vol 33 no 18 pp 222ndash248 2007

[27] B Huang and G Cao ldquoThe Path Planning Research forMobile Robot Based on the Artificial Potential FieldrdquoComputerEngineering and Applications vol 27 pp 26ndash28 2006

[28] K Jung J Kim and T Jeon ldquoCollision Avoidance of MultiplePath-planning using Fuzzy Inference Systemrdquo in Proceedings ofKIIS Spring Conference vol 19 pp 278ndash288 2009

[29] Q ZHU ldquoAnt Algorithm for Navigation of Multi-Robot Move-ment in Unknown Environmentrdquo Journal of Software vol 17no 9 p 1890 2006

[30] J Borenstein and Y Koren ldquoThe vector field histogrammdashfastobstacle avoidance for mobile robotsrdquo IEEE Transactions onRobotics and Automation vol 7 no 3 pp 278ndash288 1991

[31] D Fox W Burgard and S Thrun ldquoThe dynamic windowapproach to collision avoidancerdquo IEEERobotics andAutomationMagazine vol 4 no 1 pp 23ndash33 1997

[32] S G Tzafestas M P Tzamtzi and G G Rigatos ldquoRobustmotion planning and control of mobile robots for collisionavoidance in terrains with moving objectsrdquo Mathematics andComputers in Simulation vol 59 no 4 pp 279ndash292 2002

[33] ldquoVision-based obstacle avoidance for a small low-cost robotrdquo inProceedings of the 4th International Conference on Informatics inControl Automation and Robotics pp 275ndash279 Angers FranceMay 2007

[34] O Khatib ldquoReal-time obstacle avoida nce for manipulators andmobile robotsrdquo International Journal of Robotics Research vol5 no 1 pp 90ndash98 1986

[35] M Tounsi and J F Le Corre ldquoTrajectory generation for mobilerobotsrdquo Mathematics and Computers in Simulation vol 41 no3-4 pp 367ndash376 1996

[36] A AAhmed T Y Abdalla and A A Abed ldquoPath Planningof Mobile Robot by using Modified Optimized Potential FieldMethodrdquo International Journal of Computer Applications vol113 no 4 pp 6ndash10 2015

[37] D N Nia H S Tang B Karasfi O R Motlagh and AC Kit ldquoVirtual force field algorithm for a behaviour-basedautonomous robot in unknown environmentsrdquo Proceedings ofthe Institution ofMechanical Engineers Part I Journal of Systemsand Control Engineering vol 225 no 1 pp 51ndash62 2011

[38] Y Wang D Mulvaney and I Sillitoe ldquoGenetic-based mobilerobot path planning using vertex heuristicsrdquo in Proceedings ofthe Proc Conf on Cybernetics Intelligent Systems p 1 2006

[39] J Silvestre-Blanes ldquoReal-time obstacle avoidance using poten-tial field for a nonholonomic vehiclerdquo Factory AutomationInTech 2010

[40] B Kovacs G Szayer F Tajti M Burdelis and P KorondildquoA novel potential field method for path planning of mobilerobots by adapting animal motion attributesrdquo Robotics andAutonomous Systems vol 82 pp 24ndash34 2016

[41] H Bing L Gang G Jiang W Hong N Nan and L Yan ldquoAroute planning method based on improved artificial potentialfield algorithmrdquo inProceedings of the 2011 IEEE 3rd InternationalConference on Communication Software and Networks (ICCSN)pp 550ndash554 Xirsquoan China May 2011

[42] P Vadakkepat Kay Chen Tan and Wang Ming-LiangldquoEvolutionary artificial potential fields and their applicationin real time robot path planningrdquo in Proceedings of the 2000Congress on Evolutionary Computation pp 256ndash263 La JollaCA USA

[43] Kim Min-Ho Heo Jung-Hun Wei Yuanlong and Lee Min-Cheol ldquoA path planning algorithm using artificial potentialfield based on probability maprdquo in Proceedings of the 2011 8thInternational Conference on Ubiquitous Robots and AmbientIntelligence (URAI 2011) pp 41ndash43 Incheon November 2011

[44] L Barnes M Fields and K Valavanis ldquoUnmanned groundvehicle swarm formation control using potential fieldsrdquo inProceedings of the Proc of the 15th IEEE Mediterranean Conf onControl Automation p 1 Athens Greece 2007

[45] D Huang L Heutte and M Loog Advanced Intelligent Com-putingTheories and Applications With Aspects of ContemporaryIntelligent Computing Techniques vol 2 Springer Berlin Heidel-berg Berlin Heidelberg 2007

[46] S S Ge and Y J Cui ldquoDynamic motion planning for mobilerobots using potential field methodrdquo Autonomous Robots vol13 no 3 pp 207ndash222 2002

[47] J-Y Zhang Z-P Zhao and D Liu ldquoPath planning method formobile robot based on artificial potential fieldrdquo Harbin GongyeDaxue XuebaoJournal of Harbin Institute of Technology vol 38no 8 pp 1306ndash1309 2006

[48] F Chen P Di J Huang H Sasaki and T Fukuda ldquoEvo-lutionary artificial potential field method based manipulatorpath planning for safe robotic assemblyrdquo in Proceedings of the20th Anniversary MHS 2009 and Micro-Nano Global COE -2009 International Symposium onMicro-NanoMechatronics andHuman Science pp 92ndash97 Japan November 2009

[49] W Su R Meng and C Yu ldquoA study on soccer robot pathplanning with fuzzy artificial potential fieldrdquo in Proceedingsof the 1st International Conference on Computing Control andIndustrial Engineering CCIE 2010 pp 386ndash390 China June2010

[50] T Weerakoon K Ishii and A A Nassiraei ldquoDead-lock freemobile robot navigation using modified artificial potentialfieldrdquo in Proceedings of the 2014 Joint 7th International Confer-ence on Soft Computing and Intelligent Systems (SCIS) and 15thInternational Symposium on Advanced Intelligent Systems (ISIS)pp 259ndash264 Kita-Kyushu Japan December 2014

[51] Z Xu R Hess and K Schilling ldquoConstraints of PotentialField for Obstacle Avoidance on Car-like Mobile Robotsrdquo IFACProceedings Volumes vol 45 no 4 pp 169ndash175 2012

[52] T P Nascimento A G Conceicao and A P Moreira ldquoMulti-Robot Systems Formation Control with Obstacle AvoidancerdquoIFAC Proceedings Volumes vol 47 no 3 pp 5703ndash5708 2014

[53] K Souhila and A Karim ldquoOptical flow based robot obstacleavoidancerdquo International Journal of Advanced Robotic Systemsvol 4 no 1 pp 13ndash16 2007

[54] J Kim and Y Do ldquoMoving Obstacle Avoidance of a MobileRobot Using a Single Camerardquo Procedia Engineering vol 41 pp911ndash916 2012

[55] S Lenseir and M Veloso ldquoVisual sonar fast obstacle avoidanceusing monocular visionrdquo in Proceedings of the 2003 IEEERSJInternational Conference on Intelligent Robots and Systems(IROS 2003) (Cat No03CH37453) pp 886ndash891 Las VegasNevada USA


[56] L Kovacs ldquoVisual Monocular Obstacle Avoidance for SmallUnmanned Vehiclesrdquo in Proceedings of the 29th IEEE Confer-ence on Computer Vision and Pattern Recognition WorkshopsCVPRW 2016 pp 877ndash884 USA July 2016

[57] L Kovacs and T Sziranyi ldquoFocus area extraction by blinddeconvolution for defining regions of interestrdquo IEEE Transac-tions on Pattern Analysis andMachine Intelligence vol 29 no 6pp 1080ndash1085 2007

[58] L Kovacs ldquoSingle image visual obstacle avoidance for lowpower mobile sensingrdquo in Advanced Concepts for IntelligentVision Systems vol 9386 of Lecture Notes in Comput Sci pp261ndash272 Springer Cham 2015

[59] E Masehian and Y Katebi ldquoSensor-based motion planning ofwheeled mobile robots in unknown dynamic environmentsrdquoJournal of Intelligent amp Robotic Systems vol 74 no 3-4 pp 893ndash914 2014

[60] Li Wang Lijun Zhao Guanglei Huo et al ldquoVisual SemanticNavigation Based onDeep Learning for IndoorMobile RobotsrdquoComplexity vol 2018 pp 1ndash12 2018

[61] V Gavrilut A Tiponut A Gacsadi and L Tepelea ldquoWall-followingMethod for an AutonomousMobile Robot using TwoIR Sensorsrdquo in Proceedings of the 12th WSEAS InternationalConference on Systems pp 22ndash24 Heraklion Greece 2008

[62] A SMatveev CWang andA V Savkin ldquoReal-time navigationof mobile robots in problems of border patrolling and avoidingcollisions with moving and deforming obstaclesrdquo Robotics andAutonomous Systems vol 60 no 6 pp 769ndash788 2012

[63] R Solea and D Cernega ldquoSliding Mode Control for Trajec-tory Tracking Problem - Performance Evaluationrdquo in ArtificialNeural Networks ndash ICANN 2009 vol 5769 of Lecture Notesin Computer Science pp 865ndash874 Springer Berlin HeidelbergBerlin Heidelberg 2009

[64] R Solea and D C Cernega ldquoObstacle Avoidance for TrajectoryTracking Control ofWheeledMobile Robotsrdquo IFAC ProceedingsVolumes vol 45 no 6 pp 906ndash911 2012

[65] J Lwowski L Sun R Mexquitic-Saavedra R Sharma andD Pack ldquoA Reactive Bearing Angle Only Obstacle Avoid-ance Technique for Unmanned Ground Vehiclesrdquo Journal ofAutomation and Control Research 2014

[66] S H Tang and C K Ang ldquoA Reactive Collision AvoidanceApproach to Mobile Robot in Dynamic Environmentrdquo Journalof Automation and Control Engineering vol 1 pp 16ndash20 2013

[67] T Bandyophadyay L Sarcione and F S Hover ldquoA simplereactive obstacle avoidance algorithm and its application inSingapore harborrdquo Springer Tracts in Advanced Robotics vol 62pp 455ndash465 2010

[68] A V Savkin and C Wang ldquoSeeking a path through thecrowd Robot navigation in unknown dynamic environmentswith moving obstacles based on an integrated environmentrepresentationrdquo Robotics and Autonomous Systems vol 62 no10 pp 1568ndash1580 2014

[69] L Adouane A Benzerrouk and P Martinet ldquoMobile robotnavigation in cluttered environment using reactive elliptictrajectoriesrdquo in Proceedings of the 18th IFACWorld Congress pp13801ndash13806 Milano Italy September 2011

[70] P Ogren and N E Leonard ldquoA convergent dynamic windowapproach to obstacle avoidancerdquo IEEE Transactions on Roboticsvol 21 no 2 pp 188ndash195 2005

[71] P Ogren and N E Leonard ldquoA tractable convergent dynamicwindow approach to obstacle avoidancerdquo in Proceedings of the2002 IEEERSJ International Conference on Intelligent Robotsand Systems pp 595ndash600 Switzerland October 2002

[72] H Zhang L Dou H Fang and J Chen ldquoAutonomous indoorexploration of mobile robots based on door-guidance andimproved dynamic window approachrdquo in Proceedings of the2009 IEEE International Conference on Robotics and Biomimet-ics ROBIO 2009 pp 408ndash413 China December 2009

[73] F P Vista A M Singh D-J Lee and K T Chong ldquoDesignconvergent Dynamic Window Approach for quadrotor nav-igationrdquo International Journal of Precision Engineering andManufacturing vol 15 no 10 pp 2177ndash2184 2014

[74] H Berti A D Sappa and O E Agamennoni ldquoImproveddynamic window approach by using lyapunov stability criteriardquoLatin American Applied Research vol 38 no 4 pp 289ndash2982008

[75] X Li F Liu J Liu and S Liang ldquoObstacle avoidance for mobilerobot based on improved dynamic window approachrdquo TurkishJournal of Electrical Engineering amp Computer Sciences vol 25no 2 pp 666ndash676 2017

[76] S M LaValle H H Gonzalez-Banos C Becker and J-CLatombe ldquoMotion strategies for maintaining visibility of amoving targetrdquo in Proceedings of the 1997 IEEE InternationalConference on Robotics and Automation ICRA Part 3 (of 4) pp731ndash736 April 1997

[77] M Adbellatif and O A Montasser ldquoUsing Ultrasonic RangeSensors to Control a Mobile Robot in Obstacle AvoidanceBehaviorrdquo in Proceedings of the In Proceeding of The SCI2001World Conference on Systemics Cybernetics and Informatics pp78ndash83 2001

[78] I Ullah F Ullah Q Ullah and S Shin ldquoSensor-Based RoboticModel for Vehicle Accident Avoidancerdquo Journal of Computa-tional Intelligence and Electronic Systems vol 1 no 1 pp 57ndash622012

[79] I Ullah F Ullah and Q Ullah ldquoA sensor based robotic modelfor vehicle collision reductionrdquo in Proceedings of the 2011 Inter-national Conference on Computer Networks and InformationTechnology (ICCNIT) pp 251ndash255 Abbottabad Pakistan July2011

[80] N Kumar Z Vamossy and Z M Szabo-Resch ldquoRobot obstacleavoidance using bumper eventrdquo in Proceedings of the 2016IEEE 11th International Symposium on Applied ComputationalIntelligence and Informatics (SACI) pp 485ndash490 TimisoaraRomania May 2016

[81] F Kunwar F Wong R B Mrad and B Benhabib ldquoGuidance-based on-line robot motion planning for the interception ofmobile targets in dynamic environmentsrdquo Journal of Intelligentamp Robotic Systems vol 47 no 4 pp 341ndash360 2006

[82] W Chih-Hung H Wei-Zhon and P Shing-Tai ldquoPerformanceEvaluation of Extended Kalman Filtering for Obstacle Avoid-ance ofMobile Robotsrdquo in Proceedings of the InternationalMultiConference of Engineers and Computer Scientist vol 1 2015

[83] C C Mendes and D F Wolf ldquoStereo-Based Autonomous Nav-igation and Obstacle Avoidancerdquo IFAC Proceedings Volumesvol 46 no 10 pp 211ndash216 2013

[84] E Owen and L Montano ldquoA robocentric motion planner fordynamic environments using the velocity spacerdquo in Proceedingsof the 2006 IEEERSJ International Conference on IntelligentRobots and Systems IROS 2006 pp 4368ndash4374 China October2006

[85] H Dong W Li J Zhu and S Duan ldquoThe Path Planning forMobile Robot Based on Voronoi Diagramrdquo in Proceedings of thein 3rd Intl Conf on IntelligentNetworks Intelligent Syst Shenyang2010


[86] Y Ho and J Liu ldquoCollision-free curvature-bounded smoothpath planning using composite Bezier curve based on Voronoidiagramrdquo in Proceedings of the 2009 IEEE International Sympo-sium on Computational Intelligence in Robotics and Automation- (CIRA 2009) pp 463ndash468Daejeon Korea (South) December2009

[87] P Qu J Xue L Ma and C Ma ldquoA constrained VFH algorithmfor motion planning of autonomous vehiclesrdquo in Proceedings ofthe IEEE Intelligent Vehicles Symposium IV 2015 pp 700ndash705Republic of Korea July 2015

[88] K Lee J C Koo H R Choi and H Moon ldquoAn RRTlowast pathplanning for kinematically constrained hyper-redundant inpiperobotrdquo in Proceedings of the 12th International Conference onUbiquitous Robots and Ambient Intelligence URAI 2015 pp 121ndash128 Republic of Korea October 2015

[89] S Kamarry L Molina E A N Carvalho and E O FreireldquoCompact RRT ANewApproach for Guided Sampling Appliedto Environment Representation and Path Planning in MobileRoboticsrdquo in Proceedings of the 12th LARS Latin AmericanRobotics Symposiumand 3rd SBRBrazilian Robotics SymposiumLARS-SBR 2015 pp 259ndash264 Brazil October 2015

[90] M Srinivasan Ramanagopal A P Nguyen and J Le Ny ldquoAMotion Planning Strategy for the Active Vision-BasedMappingof Ground-Level Structuresrdquo IEEE Transactions on AutomationScience and Engineering vol 15 no 1 pp 356ndash368 2018

[91] C Fulgenzi A Spalanzani and C Laugier ldquoDynamic obstacleavoidance in uncertain environment combining PVOs andoccupancy gridrdquo in Proceedings of the IEEE International Con-ference on Robotics and Automation (ICRA rsquo07) pp 1610ndash1616April 2007

[92] Y Wang A Goila R Shetty M Heydari A Desai and HYang ldquoObstacle Avoidance Strategy and Implementation forUnmanned Ground Vehicle Using LIDARrdquo SAE InternationalJournal of Commercial Vehicles vol 10 no 1 pp 50ndash55 2017

[93] R Lagisetty N K Philip R Padhi and M S Bhat ldquoObjectdetection and obstacle avoidance for mobile robot using stereocamerardquo in Proceedings of the IEEE International Conferenceon Control Applications pp 605ndash610 Hyderabad India August2013

[94] J Michels A Saxena and A Y Ng ldquoHigh speed obstacleavoidance usingmonocular vision and reinforcement learningrdquoin Proceedings of the ICML 2005 22nd International Conferenceon Machine Learning pp 593ndash600 Germany August 2005

[95] H Seraji and A Howard ldquoBehavior-based robot navigation onchallenging terrain A fuzzy logic approachrdquo IEEE Transactionson Robotics and Automation vol 18 no 3 pp 308ndash321 2002

[96] S Hrabar ldquo3D path planning and stereo-based obstacle avoid-ance for rotorcraft UAVsrdquo in Proceedings of the 2008 IEEERSJInternational Conference on Intelligent Robots and SystemsIROS pp 807ndash814 France September 2008

[97] S H Han Na and H Jeong ldquoStereo-based road obstacledetection and trackingrdquo in Proceedings of the In 13th Int Confon ICACT IEEE pp 1181ndash1184 2011

[98] D N Bhat and S K Nayar ldquoStereo and Specular ReflectionrdquoInternational Journal of Computer Vision vol 26 no 2 pp 91ndash106 1998

[99] P S Sharma and D N G Chitaliya ldquoObstacle Avoidance UsingStereo Vision A Surveyrdquo International Journal of InnovativeResearch in Computer and Communication Engineering vol 03no 01 pp 24ndash29 2015

[100] A Typiak ldquoUse of laser rangefinder to detecting in surround-ings of mobile robot the obstaclesrdquo in Proceedings of the The

25th International Symposium on Automation and Robotics inConstruction pp 246ndash251 Vilnius Lithuania June 2008

[101] P D Baldoni Y Yang and S Kim ldquoDevelopment of EfficientObstacle Avoidance for aMobile Robot Using Fuzzy Petri Netsrdquoin Proceedings of the 2016 IEEE 17th International Conferenceon Information Reuse and Integration (IRI) pp 265ndash269 Pitts-burgh PA USA July 2016

[102] D Janglova ldquoNeural Networks in Mobile Robot MotionrdquoInternational Journal of Advanced Robotic Systems vol 1 no 1p 2 2004

[103] D Kiss and G Tevesz ldquoAdvanced dynamic window based navi-gation approach using model predictive controlrdquo in Proceedingsof the 2012 17th International Conference on Methods amp Modelsin Automation amp Robotics (MMAR) pp 148ndash153 MiedzyzdrojePoland August 2012

[104] P Saranrittichai N Niparnan and A Sudsang ldquoRobust localobstacle avoidance for mobile robot based on Dynamic Win-dow approachrdquo in Proceedings of the 2013 10th InternationalConference on Electrical EngineeringElectronics ComputerTelecommunications and Information Technology (ECTI-CON2013) pp 1ndash4 Krabi Thailand May 2013

[105] C H Hommes ldquoHeterogeneous agent models in economicsand financerdquo in Handbook of Computational Economics vol 2pp 1109ndash1186 Elsevier 2006

[106] L Chengqing M H Ang H Krishnan and L S Yong ldquoVirtualObstacle Concept for Local-minimum-recovery in Potential-field Based Navigationrdquo in Proceedings of the IEEE Int Conf onRobotics Automation pp 983ndash988 San Francisco 2000

[107] Y Koren and J Borenstein ldquoPotential field methods and theirinherent limitations formobile robot navigationrdquo inProceedingsof the IEEE International Conference on Robotics and Automa-tion pp 1398ndash1404 April 1991

[108] S S Ge and Y J Cui ldquoNew potential functions for mobile robotpath planningrdquo IEEE Transactions on Robotics and Automationvol 16 no 5 pp 615ndash620 2000

[109] Q Jia and X Wang ldquoAn improved potential field method forpath planningrdquo in Proceedings of the Chinese Control and Deci-sion Conference (CCDC rsquo10) pp 2265ndash2270 Xuzhou ChinaMay 2010

[110] J Canny and M Lin ldquoAn opportunistic global path plannerrdquoin Proceedings of the IEEE International Conference on Roboticsand Automation pp 1554ndash1559 Cincinnati OH USA

[111] K-Y Chen P A Lindsay P J Robinson and H A AbbassldquoA hierarchical conflict resolution method for multi-agentpath planningrdquo in Proceedings of the 2009 IEEE Congress onEvolutionary Computation CEC 2009 pp 1169ndash1176 NorwayMay 2009

[112] Y Morales A Carballo E Takeuchi A Aburadani and TTsubouchi ldquoAutonomous robot navigation in outdoor clutteredpedestrian walkwaysrdquo Journal of Field Robotics vol 26 no 8pp 609ndash635 2009

[113] D Kim J Kim J Bae and Y Soh ldquoDevelopment of anEnhanced Obstacle Avoidance Algorithm for a Network-BasedAutonomous Mobile Robotrdquo in Proceedings of the 2010 Inter-national Conference on Intelligent Computation Technology andAutomation (ICICTA) pp 102ndash105 Changsha ChinaMay 2010

[114] V I Utkin ldquoSlidingmode controlrdquo inVariable structure systemsfrom principles to implementation vol 66 of IEE Control EngSer pp 3ndash17 IEE London 2004

[115] N Siddique and H Adeli ldquoNature inspired computing anoverview and some future directionsrdquo Cognitive Computationvol 7 no 6 pp 706ndash714 2015


[116] MH Saffari andM JMahjoob ldquoBee colony algorithm for real-time optimal path planning of mobile robotsrdquo in Proceedings ofthe 5th International Conference on Soft Computing Computingwith Words and Perceptions in System Analysis Decision andControl (ICSCCW rsquo09) pp 1ndash4 IEEE September 2009

[117] L Na and F Zuren ldquoNumerical potential field and ant colonyoptimization based path planning in dynamic environmentrdquo inProceedings of the 6th World Congress on Intelligent Control andAutomation WCICA 2006 pp 8966ndash8970 China June 2006

[118] C A Floudas Deterministic global optimization vol 37 ofNonconvex Optimization and its Applications Kluwer AcademicPublishers Dordrecht 2000

[119] J-H Lin and L-R Huang ldquoChaotic Bee Swarm OptimizationAlgorithm for Path Planning of Mobile Robotsrdquo in Proceedingsof the 10th WSEAS International Conference on EvolutionaryComputing Prague Czech Republic 2009

[120] L S Sierakowski and C A Coelho ldquoBacteria ColonyApproaches with Variable Velocity Applied to Path Optimiza-tion of Mobile Robotsrdquo in Proceedings of the 18th InternationalCongress of Mechanical Engineering Ouro Preto MG Brazil2005

[121] C A Sierakowski and L D S Coelho ldquoStudy of Two SwarmIntelligence Techniques for Path Planning of Mobile Robots16thIFAC World Congressrdquo Prague 2005

[122] N HadiAbbas and F Mahdi Ali ldquoPath Planning of anAutonomous Mobile Robot using Directed Artificial BeeColony Algorithmrdquo International Journal of Computer Applica-tions vol 96 no 11 pp 11ndash16 2014

[123] H Chen Y Zhu and K Hu ldquoAdaptive bacterial foragingoptimizationrdquo Abstract and Applied Analysis Art ID 108269 27pages 2011

[124] X Yang Nature-Inspired Metaheuristic Algorithms LuniverPress United Kingdom 2 edition 2010

[125] D K Chaturvedi ldquoSoft Computing Techniques and TheirApplicationsrdquo in Mathematical Models Methods and Applica-tions Industrial and Applied Mathematics pp 31ndash40 SpringerSingapore Singapore 2015

[126] N B Hui and D K Pratihar ldquoA comparative study on somenavigation schemes of a real robot tackling moving obstaclesrdquoRobotics and Computer-IntegratedManufacturing vol 25 no 4-5 pp 810ndash828 2009

[127] F J Pelletier ldquoReview of Mathematics of Fuzzy Logicsrdquo TheBulleting ofn Symbolic Logic vol 6 no 3 pp 342ndash346 2000

[128] L A Zadeh ldquoFuzzy setsrdquo Information and Computation vol 8pp 338ndash353 1965

[129] R Ramanathan ldquoService-Driven Computingrdquo in Handbook ofResearch on Architectural Trends in Service-Driven ComputingAdvances in Systems Analysis Software Engineering and HighPerformance Computing pp 1ndash25 IGI Global 2014

[130] F Cupertino V Giordano D Naso and L Delfine ldquoFuzzycontrol of a mobile robotrdquo IEEE Robotics and AutomationMagazine vol 13 no 4 pp 74ndash81 2006

[131] H A Hagras ldquoA hierarchical type-2 fuzzy logic control archi-tecture for autonomous mobile robotsrdquo IEEE Transactions onFuzzy Systems vol 12 no 4 pp 524ndash539 2004

[132] K P Valavanis L Doitsidis M Long and R RMurphy ldquoA casestudy of fuzzy-logic-based robot navigationrdquo IEEE Robotics andAutomation Magazine vol 13 no 3 pp 93ndash107 2006

[133] RuiWangMingWang YongGuan andXiaojuan Li ldquoModelingand Analysis of the Obstacle-Avoidance Strategies for a MobileRobot in a Dynamic Environmentrdquo Mathematical Problems inEngineering vol 2015 pp 1ndash11 2015

[134] J Vascak and M Pala ldquoAdaptation of fuzzy cognitive mapsfor navigation purposes bymigration algorithmsrdquo InternationalJournal of Artificial Intelligence vol 8 pp 20ndash37 2012

[135] M J Er and C Deng ldquoOnline tuning of fuzzy inferencesystems using dynamic fuzzyQ-learningrdquo IEEE Transactions onSystems Man and Cybernetics Part B Cybernetics vol 34 no3 pp 1478ndash1489 2004

[136] X Yang M Moallem and R V Patel ldquoA layered goal-orientedfuzzy motion planning strategy for mobile robot navigationrdquoIEEE Transactions on Systems Man and Cybernetics Part BCybernetics vol 35 no 6 pp 1214ndash1224 2005

[137] F Abdessemed K Benmahammed and E Monacelli ldquoA fuzzy-based reactive controller for a non-holonomic mobile robotrdquoRobotics andAutonomous Systems vol 47 no 1 pp 31ndash46 2004

[138] M Shayestegan and M H Marhaban ldquoMobile robot safenavigation in unknown environmentrdquo in Proceedings of the 2012IEEE International Conference on Control System Computingand Engineering ICCSCE 2012 pp 44ndash49 Malaysia November2012

[139] X Li and B-J Choi ldquoDesign of obstacle avoidance system formobile robot using fuzzy logic systemsrdquo International Journalof Smart Home vol 7 no 3 pp 321ndash328 2013

[140] Y Chang R Huang and Y Chang ldquoA Simple Fuzzy MotionPlanning Strategy for Autonomous Mobile Robotsrdquo in Pro-ceedings of the IECON 2007 - 33rd Annual Conference of theIEEE Industrial Electronics Society pp 477ndash482 Taipei TaiwanNovember 2007

[141] D R Parhi and M K Singh ldquoIntelligent fuzzy interfacetechnique for the control of an autonomous mobile robotrdquoProceedings of the Institution of Mechanical Engineers Part CJournal of Mechanical Engineering Science vol 222 no 11 pp2281ndash2292 2008

[142] Y Qin F Da-Wei HWang andW Li ldquoFuzzy Control ObstacleAvoidance for Mobile Robotrdquo Journal of Hebei University ofTechnology 2007

[143] H-H Lin and C-C Tsai ldquoLaser pose estimation and trackingusing fuzzy extended information filtering for an autonomousmobile robotrdquo Journal of Intelligent amp Robotic Systems vol 53no 2 pp 119ndash143 2008

[144] S Parasuraman ldquoSensor fusion for mobile robot navigationFuzzy Associative Memoryrdquo in Proceedings of the 2nd Interna-tional Symposium on Robotics and Intelligent Sensors 2012 IRIS2012 pp 251ndash256 Malaysia September 2012

[145] M Dupre and S X Yang ldquoTwo-stage fuzzy logic-based con-troller for mobile robot navigationrdquo in Proceedings of the 2006IEEE International Conference on Mechatronics and Automa-tion ICMA 2006 pp 745ndash750 China June 2006

[146] S Nurmaini and S Z M Hashim ldquoAn embedded fuzzytype-2 controller based sensor behavior for mobile robotrdquo inProceedings of the 8th International Conference on IntelligentSystems Design and Applications ISDA 2008 pp 29ndash34 TaiwanNovember 2008

[147] M F Selekwa D D Dunlap D Shi and E G Collins Jr ldquoRobotnavigation in very cluttered environments by preference-basedfuzzy behaviorsrdquo Robotics and Autonomous Systems vol 56 no3 pp 231ndash246 2008

[148] U Farooq K M Hasan A Raza M Amar S Khan and SJavaid ldquoA low cost microcontroller implementation of fuzzylogic based hurdle avoidance controller for a mobile robotrdquo inProceedings of the 2010 3rd IEEE International Conference onComputer Science and Information Technology (ICCSIT 2010)pp 480ndash485 Chengdu China July 2010


[149] Fahmizal and C-H Kuo ldquoTrajectory and heading trackingof a mecanum wheeled robot using fuzzy logic controlrdquo inProceedings of the 2016 International Conference on Instrumenta-tion Control and Automation ICA 2016 pp 54ndash59 IndonesiaAugust 2016

[150] S H A Mohammad M A Jeffril and N Sariff ldquoMobilerobot obstacle avoidance by using Fuzzy Logic techniquerdquo inProceedings of the 2013 IEEE 3rd International Conference onSystem Engineering and Technology ICSET 2013 pp 331ndash335Malaysia August 2013

[151] J Berisha and A Shala ldquoApplication of Fuzzy Logic Controllerfor Obstaclerdquo in Proceedings of the 5th Mediterranean Conf onEmbedded Comp pp 200ndash205 Bar Montenegro 2016

[152] MM Almasri KM Elleithy andAM Alajlan ldquoDevelopmentof efficient obstacle avoidance and line following mobile robotwith the integration of fuzzy logic system in static and dynamicenvironmentsrdquo in Proceedings of the IEEE Long Island SystemsApplications and Technology Conference LISAT 2016 USA

[153] M I Ibrahim N Sariff J Johari and N Buniyamin ldquoMobilerobot obstacle avoidance in various type of static environ-ments using fuzzy logic approachrdquo in Proceedings of the 2014International Conference on Electrical Electronics and SystemEngineering (ICEESE) pp 83ndash88 Kuala Lumpur December2014

[154] A Al-Mayyahi and W Wang ldquoFuzzy inference approach forautonomous ground vehicle navigation in dynamic environ-mentrdquo in Proceedings of the 2014 IEEE International Conferenceon Control System Computing and Engineering (ICCSCE) pp29ndash34 Penang Malaysia November 2014

[155] U Farooq M Amar E Ul Haq M U Asad and H MAtiq ldquoMicrocontroller based neural network controlled lowcost autonomous vehiclerdquo in Proceedings of the 2010 The 2ndInternational Conference on Machine Learning and ComputingICMLC 2010 pp 96ndash100 India February 2010

[156] U Farooq M Amar K M Hasan M K Akhtar M U Asadand A Iqbal ldquoA low cost microcontroller implementation ofneural network based hurdle avoidance controller for a car-likerobotrdquo in Proceedings of the 2nd International Conference onComputer and Automation Engineering ICCAE 2010 pp 592ndash597 Singapore February 2010

[157] K-HChi andM-F R Lee ldquoObstacle avoidance inmobile robotusing neural networkrdquo in Proceedings of the 2011 InternationalConference on Consumer Electronics Communications and Net-works CECNet rsquo11 pp 5082ndash5085 April 2011

[158] V Ganapathy S C Yun and H K Joe ldquoNeural Q-Iearningcontroller for mobile robotrdquo in Proceedings of the IEEElASMEInt Conf on Advanced Intelligent Mechatronics pp 863ndash8682009

[159] S J Yoo ldquoAdaptive neural tracking and obstacle avoidance ofuncertain mobile robots with unknown skidding and slippingrdquoInformation Sciences vol 238 pp 176ndash189 2013

[160] J Qiao Z Hou and X Ruan ldquoApplication of reinforcementlearning based on neural network to dynamic obstacle avoid-ancerdquo in Proceedings of the 2008 IEEE International Conferenceon Information andAutomation ICIA 2008 pp 784ndash788ChinaJune 2008

[161] A Chohra C Benmehrez and A Farah ldquoNeural NavigationApproach for Intelligent Autonomous Vehicles (IAV) in Par-tially Structured Environmentsrdquo Applied Intelligence vol 8 no3 pp 219ndash233 1998

[162] R Glasius A Komoda and S C AM Gielen ldquoNeural networkdynamics for path planning and obstacle avoidancerdquo NeuralNetworks vol 8 no 1 pp 125ndash133 1995

[163] VGanapathy C Y Soh and J Ng ldquoFuzzy and neural controllersfor acute obstacle avoidance in mobile robot navigationrdquo inProceedings of the IEEEASME International Conference onAdvanced IntelligentMechatronics (AIM rsquo09) pp 1236ndash1241 July2009

[164] Guo-ShengYang Er-KuiChen and Cheng-WanAn ldquoMobilerobot navigation using neural Q-learningrdquo in Proceedings ofthe 2004 International Conference on Machine Learning andCybernetics pp 48ndash52 Shanghai China

[165] C Li J Zhang and Y Li ldquoApplication of Artificial NeuralNetwork Based onQ-learning forMobile Robot Path Planningrdquoin Proceedings of the 2006 IEEE International Conference onInformation Acquisition pp 978ndash982 Veihai China August2006

[166] K Macek I Petrovic and N Peric ldquoA reinforcement learningapproach to obstacle avoidance ofmobile robotsrdquo inProceedingsof the 7th International Workshop on Advanced Motion Control(AMC rsquo02) pp 462ndash466 IEEE Maribor Slovenia July 2002

[167] R Akkaya O Aydogdu and S Canan ldquoAn ANN Based NARXGPSDR System for Mobile Robot Positioning and ObstacleAvoidancerdquo Journal of Automation and Control vol 1 no 1 p13 2013

[168] P K Panigrahi S Ghosh and D R Parhi ldquoA novel intelligentmobile robot navigation technique for avoiding obstacles usingRBF neural networkrdquo in Proceedings of the 2014 InternationalConference on Control Instrumentation Energy and Communi-cation (CIEC) pp 1ndash6 Calcutta India January 2014

[169] U Farooq M Amar M U Asad A Hanif and S O SaleHldquoDesign and Implementation of Neural Network Basedrdquo vol 6pp 83ndash89 2014

[170] AMedina-Santiago J L Camas-Anzueto J A Vazquez-FeijooH R Hernandez-De Leon and R Mota-Grajales ldquoNeuralcontrol system in obstacle avoidance in mobile robots usingultrasonic sensorsrdquo Journal of Applied Research and Technologyvol 12 no 1 pp 104ndash110 2014

[171] U A Syed F Kunwar and M Iqbal ldquoGuided autowave pulsecoupled neural network (GAPCNN) based real time pathplanning and an obstacle avoidance scheme for mobile robotsrdquoRobotics and Autonomous Systems vol 62 no 4 pp 474ndash4862014

[172] HQu S X Yang A RWillms andZ Yi ldquoReal-time robot pathplanning based on a modified pulse-coupled neural networkmodelrdquo IEEE Transactions on Neural Networks and LearningSystems vol 20 no 11 pp 1724ndash1739 2009

[173] M Pfeiffer M Schaeuble J Nieto R Siegwart and C CadenaldquoFrom perception to decision A data-driven approach toend-to-end motion planning for autonomous ground robotsrdquoin Proceedings of the 2017 IEEE International Conference onRobotics and Automation ICRA 2017 pp 1527ndash1533 SingaporeJune 2017

[174] D FOGEL ldquoAn introduction to genetic algorithms MelanieMitchell MIT Press Cambridge MA 1996 000 (cloth) 270pprdquo Bulletin of Mathematical Biology vol 59 no 1 pp 199ndash2041997

[175] Tu Jianping and S Yang ldquoGenetic algorithm based path plan-ning for amobile robotrdquo in Proceedings of the IEEE InternationalConference on Robotics and Automation IEEE ICRA 2003 pp1221ndash1226 Taipei Taiwan


[176] Y Zhou L Zheng andY Li ldquoAn improved genetic algorithm formobile robotic path planningrdquo in Proceedings of the 2012 24thChinese Control andDecision Conference CCDC 2012 pp 3255ndash3260 China May 2012

[177] K H Sedighi K Ashenayi T W Manikas R L Wainwrightand H-M Tai ldquoAutonomous local path planning for a mobilerobot using a genetic algorithmrdquo in Proceedings of the 2004Congress on Evolutionary Computation CEC2004 pp 1338ndash1345 USA June 2004

[178] A Tuncer and M Yildirim ldquoDynamic path planning of mobilerobots with improved genetic algorithmrdquo Computers and Elec-trical Engineering vol 38 no 6 pp 1564ndash1572 2012

[179] I Chaari A Koubaa H Bennaceur S Trigui and K Al-Shalfan ldquosmartPATH A hybrid ACO-GA algorithm for robotpath planningrdquo in Proceedings of the 2012 IEEE Congress onEvolutionary Computation (CEC) pp 1ndash8 Brisbane AustraliaJune 2012

[180] R S Lim H M La and W Sheng ldquoA robotic crack inspectionand mapping system for bridge deck maintenancerdquo IEEETransactions on Automation Science and Engineering vol 11 no2 pp 367ndash378 2014

[181] T Geisler and T W Monikas ldquoAutonomous robot navigationsystem using a novel value encoded genetic algorithmrdquoMidwestSymposium onCircuits and Systems vol 3 pp III45ndashIII48 2002

[182] T Geisler and T Manikas ldquoAutonomous robot navigationsystem using a novel value encoded genetic algorithmrdquo inProceedings of the Midwest Symposium on Circuits and Systemspp III-45ndashIII-48 Tulsa OK USA

[183] P Calistri A Giovannini and Z Hubalek ldquoEpidemiology ofWest Nile in Europe and in theMediterranean basinrdquoTheOpenVirology Journal vol 4 pp 29ndash37 2010

[184] Hu Yanrong S Yang Xu Li-Zhong and M Meng ldquoA Knowl-edge Based Genetic Algorithm for Path Planning in Unstruc-tured Mobile Robot Environmentsrdquo in Proceedings of the 2004IEEE International Conference on Robotics and Biomimetics pp767ndash772 Shenyang China

[185] P Moscato On evolution search optimization genetic algo-rithms and martial arts towards memetic algorithms Cal-tech Concurrent Computation Program California Institute ofTechnology Pasadena 1989

[186] A Caponio G L Cascella F Neri N Salvatore and MSumner ldquoA fast adaptive memetic algorithm for online andoffline control design of PMSM drivesrdquo IEEE Transactions onSystems Man and Cybernetics Part B Cybernetics vol 37 no1 pp 28ndash41 2007

[187] F Neri and E Mininno ldquoMemetic compact differential evolu-tion for cartesian robot controlrdquo IEEE Computational Intelli-gence Magazine vol 5 no 2 pp 54ndash65 2010

[188] N Shahidi H Esmaeilzadeh M Abdollahi and C LucasldquoMemetic algorithm based path planning for a mobile robotrdquoin Proceedings of the int conf pp 56ndash59 2004

[189] J Botzheim Y Toda and N Kubota ldquoBacterial memeticalgorithm for offline path planning of mobile robotsrdquo MemeticComputing vol 4 no 1 pp 73ndash86 2012

[190] J Botzheim C Cabrita L T Koczy and A E Ruano ldquoFuzzyrule extraction by bacterial memetic algorithmsrdquo InternationalJournal of Intelligent Systems vol 24 no 3 pp 312ndash339 2009

[191] J Kennedy and R Eberhart ldquoParticle swarm optimizationrdquoin Proceedings of the IEEE International Conference on NeuralNetworks (ICNN rsquo95) vol 4 pp 1942ndash1948 Perth WesternAustralia November-December 1995

[192] A-M Batiha and B Batiha ldquoDifferential transformationmethod for a reliable treatment of the nonlinear biochemicalreaction modelrdquo Advanced Studies in Biology vol 3 pp 355ndash360 2011

[193] Y Tang Z Wang and J-A Fang ldquoFeedback learning particleswarm optimizationrdquo Applied Soft Computing vol 11 no 8 pp4713ndash4725 2011

[194] Y Tang Z Wang and J Fang ldquoParameters identification ofunknown delayed genetic regulatory networks by a switchingparticle swarm optimization algorithmrdquo Expert Systems withApplications vol 38 no 3 pp 2523ndash2535 2011

[195] Yong Ma M Zamirian Yadong Yang Yanmin Xu and JingZhang ldquoPath Planning for Mobile Objects in Four-DimensionBased on Particle Swarm Optimization Method with PenaltyFunctionrdquoMathematical Problems in Engineering vol 2013 pp1ndash9 2013

[196] M K Rath and B B V L Deepak ldquoPSO based systemarchitecture for path planning of mobile robot in dynamicenvironmentrdquo in Proceedings of the Global Conference on Com-munication Technologies GCCT 2015 pp 797ndash801 India April2015

[197] YMaHWang Y Xie andMGuo ldquoPath planning formultiplemobile robots under double-warehouserdquo Information Sciencesvol 278 pp 357ndash379 2014

[198] E Masehian and D Sedighizadeh ldquoMulti-objective PSO- andNPSO-based algorithms for robot path planningrdquo Advances inElectrical and Computer Engineering vol 10 no 4 pp 69ndash762010

[199] Y Zhang D-W Gong and J-H Zhang ldquoRobot path planningin uncertain environment using multi-objective particle swarmoptimizationrdquo Neurocomputing vol 103 pp 172ndash185 2013

[200] Q Li C Zhang Y Xu and Y Yin ldquoPath planning of mobilerobots based on specialized genetic algorithm and improvedparticle swarm optimizationrdquo in Proceedings of the 31st ChineseControl Conference CCC 2012 pp 7204ndash7209 China July 2012

[201] Q Zhang and S Li ldquoA global path planning approach based onparticle swarm optimization for a mobile robotrdquo in Proceedingsof the 7th WSEAS Int Conf on Robotics Control and Manufac-turing Tech pp 111ndash222 Hangzhou China 2007

[202] D-W Gong J-H Zhang and Y Zhang ldquoMulti-objective parti-cle swarm optimization for robot path planning in environmentwith danger sourcesrdquo Journal of Computers vol 6 no 8 pp1554ndash1561 2011

[203] Y Wang and X Yu ldquoResearch for the Robot Path PlanningControl Strategy Based on the Immune Particle Swarm Opti-mization Algorithmrdquo in Proceedings of the 2012 Second Interna-tional Conference on Intelligent System Design and EngineeringApplication (ISDEA) pp 724ndash727 Sanya China January 2012

[204] S Doctor G K Venayagamoorthy and V G Gudise ldquoOptimalPSO for collective robotic search applicationsrdquo in Proceedings ofthe 2004 Congress on Evolutionary Computation CEC2004 pp1390ndash1395 USA June 2004

[205] L Smith G KVenayagamoorthy and PGHolloway ldquoObstacleavoidance in collective robotic search using particle swarmoptimizationrdquo in Proceedings of the in IEEE Swarm IntelligenceSymposium Indianapolis USA 2006

[206] L Lu and D Gong ldquoRobot path planning in unknown envi-ronments using particle swarm optimizationrdquo in Proceedings ofthe 4th International Conference onNatural Computation (ICNCrsquo08) pp 422ndash426 IEEE Jinan China October 2008

[207] R Bibi B S Chowdhry andRA Shah ldquoPSObased localizationof multiple mobile robots emplying LEGO EV3rdquo in Proceedings


of the 2018 International Conference on ComputingMathematicsand Engineering Technologies (iCoMET) pp 1ndash5 SukkurMarch2018

[208] A Z Nasrollahy and H H S Javadi ldquoUsing particle swarmoptimization for robot path planning in dynamic environmentswith moving obstacles and targetrdquo in Proceedings of the UKSim3rd EuropeanModelling Symposium on ComputerModelling andSimulation EMS 2009 pp 60ndash65 Greece November 2009

[209] Hua-Qing Min Jin-Hui Zhu and Xi-Jing Zheng ldquoObsta-cle avoidance with multi-objective optimization by PSO indynamic environmentrdquo in Proceedings of 2005 InternationalConference on Machine Learning and Cybernetics pp 2950ndash2956 Vol 5 Guangzhou China August 2005

[210] Yuan-Qing Qin De-Bao Sun Ning Li and Yi-GangCen ldquoPath planning for mobile robot using the particle swarmoptimizationwithmutation operatorrdquo inProceedings of the 2004International Conference on Machine Learning and Cyberneticspp 2473ndash2478 Shanghai China

[211] B Deepak and D Parhi ldquoPSO based path planner of anautonomous mobile robotrdquo Open Computer Science vol 2 no2 2012

[212] P Curkovic and B Jerbic ldquoHoney-bees optimization algorithmapplied to path planning problemrdquo International Journal ofSimulation Modelling vol 6 no 3 pp 154ndash164 2007

[213] A Haj Darwish A Joukhadar M Kashkash and J Lam ldquoUsingthe Bees Algorithm for wheeled mobile robot path planning inan indoor dynamic environmentrdquo Cogent Engineering vol 5no 1 2018

[214] M Park J Jeon and M Lee ldquoObstacle avoidance for mobilerobots using artificial potential field approach with simulatedannealingrdquo in Proceedings of the IEEE International Symposiumon Industrial Electronics Proceedings (ISIE rsquo01) pp 1530ndash1535June 2001

[215] H Martınez-Alfaro and S Gomez-Garcıa ldquoMobile robot pathplanning and tracking using simulated annealing and fuzzylogic controlrdquo Expert Systems with Applications vol 15 no 3-4 pp 421ndash429 1998

[216] Q Zhu Y Yan and Z Xing ldquoRobot Path Planning Based onArtificial Potential Field Approach with Simulated Annealingrdquoin Proceedings of the Sixth International Conference on IntelligentSystems Design and Applications pp 622ndash627 Jian ChinaOctober 2006

[217] H Miao and Y Tian ldquoRobot path planning in dynamicenvironments using a simulated annealing based approachrdquoin Proceedings of the 2008 10th International Conference onControl Automation Robotics and Vision (ICARCV) pp 1253ndash1258 Hanoi Vietnam December 2008

[218] O Montiel-Ross R Sepulveda O Castillo and P Melin ldquoAntcolony test center for planning autonomous mobile robotnavigationrdquo Computer Applications in Engineering Educationvol 21 no 2 pp 214ndash229 2013

[219] C-F Juang C-M Lu C Lo and C-Y Wang ldquoAnt colony opti-mization algorithm for fuzzy controller design and its FPGAimplementationrdquo IEEE Transactions on Industrial Electronicsvol 55 no 3 pp 1453ndash1462 2008

[220] G-Z Tan H He and A Sloman ldquoAnt colony system algorithmfor real-time globally optimal path planning of mobile robotsrdquoActa Automatica Sinica vol 33 no 3 pp 279ndash285 2007

[221] N A Vien N H Viet S Lee and T Chung ldquoObstacleAvoidance Path Planning for Mobile Robot Based on Ant-Q Reinforcement Learning Algorithmrdquo in Advances in Neural

Networks ndash ISNN 2007 vol 4491 of Lecture Notes in ComputerScience pp 704ndash713 Springer Berlin Heidelberg Berlin Hei-delberg 2007

[222] K Ioannidis G C Sirakoulis and I Andreadis ldquoCellular antsA method to create collision free trajectories for a cooperativerobot teamrdquo Robotics and Autonomous Systems vol 59 no 2pp 113ndash127 2011

[223] B R Fajen andWHWarren ldquoBehavioral dynamics of steeringobstacle avoidance and route selectionrdquo J Exp Psychol HumPercept Perform vol 29 pp 343ndash362 2003

[224] P J Beek C E Peper and D F Stegeman ldquoDynamical modelsof movement coordinationrdquo Human Movement Science vol 14no 4-5 pp 573ndash608 1995

[225] J A S Kelso Coordination Dynamics Issues and TrendsSpringer-Verlag Berlin 2004

[226] R Fajen W H Warren S Temizer and L P Kaelbling ldquoAdynamic model of visually-guided steering obstacle avoidanceand route selectionrdquo Int J Comput Vision vol 54 pp 13ndash342003

[227] K Patanarapeelert T D Frank R Friedrich P J Beek and IM Tang ldquoTheoretical analysis of destabilization resonances intime-delayed stochastic second-order dynamical systems andsome implications for human motor controlrdquo Physical ReviewE Statistical Nonlinear and Soft Matter Physics vol 73 no 2Article ID 021901 2006

[228] G Schoner M Dose and C Engels ldquoDynamics of behaviortheory and applications for autonomous robot architecturesrdquoRobotics amp Autonomous Systems vol 16 no 2-4 pp 213ndash2451995

[229] T D Frank M J Richardson S M Lopresti-Goodman andMT Turvey ldquoOrder parameter dynamics of body-scaled hysteresisandmode transitions in grasping behaviorrdquo Journal of BiologicalPhysics vol 35 no 2 pp 127ndash147 2009

[230] H Haken Synergetic Cmputers and Cognition A Top-DownApproach to Neural Nets Springer Berlin Germany 1991

[231] T D Frank T D Gifford and S Chiangga ldquoMinimalisticmodelfor navigation of mobile robots around obstacles based oncomplex-number calculus and inspired by human navigationbehaviorrdquoMathematics andComputers in Simulation vol 97 pp108ndash122 2014

[232] J Yang P Chen H-J Rong and B Chen ldquoLeast mean p-powerextreme learning machine for obstacle avoidance of a mobilerobotrdquo in Proceedings of the 2016 International Joint Conferenceon Neural Networks IJCNN 2016 pp 1968ndash1976 Canada July2016

[233] Q Zheng and F Li ldquoA survey of scheduling optimizationalgorithms studies on automated storage and retrieval systems(ASRSs)rdquo in Proceedings of the 2010 3rd International Confer-ence on Advanced Computer Theory and Engineering (ICACTE2010) pp V1-369ndashV1-372 Chengdu China August 2010

[234] M A P Garcia O Montiel O Castillo R Sepulveda and PMelin ldquoPath planning for autonomous mobile robot navigationwith ant colony optimization and fuzzy cost function evalua-tionrdquo Applied Soft Computing vol 9 no 3 pp 1102ndash1110 2009

[235] F Neri and C Cotta ldquoMemetic algorithms and memeticcomputing optimization a literature reviewrdquo Swarm and Evo-lutionary Computation vol 2 pp 1ndash14 2012

[236] J M Hall M K Lee B Newman et al ldquoLinkage of early-onsetfamilial breast cancer to chromosome 17q21rdquo Science vol 250no 4988 pp 1684ndash1689 1990


[237] A L Bolaji A T Khader M A Al-Betar andM A AwadallahldquoArtificial bee colony algorithm its variants and applications asurveyrdquo Journal of Theoretical and Applied Information Technol-ogy vol 47 no 2 pp 434ndash459 2013

[238] S Haykin Neural Networks A Comprehensive FoundationPrentice Hall New Jersey 1999

[239] H Burchardt and R Salomon ldquoImplementation of path plan-ning using genetic algorithms on mobile robotsrdquo in Proceedingsof the IEEE Congress on Evolutionary Computation (CEC rsquo06)pp 1831ndash1836 July 2006

[240] K Su Y Wang and X Hu ldquoRobot Path Planning Based onRandom Coding Particle Swarm Optimizationrdquo InternationalJournal of Advanced Computer Science and Applications vol 6no 4 2015

[241] D Karaboga B Gorkemli C Ozturk and N Karaboga ldquoAcomprehensive survey artificial bee colony (ABC) algorithmand applicationsrdquo Artificial Intelligence Review vol 42 pp 21ndash57 2014

[242] L A Barroso and U Holzle ldquoThe case for energy-proportionalcomputingrdquo The Computer Journal vol 40 no 12 pp 33ndash372007

[243] A Abraham ldquoAdaptation of Fuzzy Inference System UsingNeural Learningrdquo in Fuzzy Systems Engineering vol 181 ofStudies in Fuzziness and Soft Computing pp 53ndash83 SpringerBerlin Heidelberg Berlin Heidelberg 2005

[244] O Obe and I Dumitrache ldquoAdaptive neuro-fuzzy controlerwith genetic training for mobile robot controlrdquo InternationalJournal of Computers Communications amp Control vol 7 no 1pp 135ndash146 2012

[245] A Zhu and S X Yang ldquoNeurofuzzy-Based Approach toMobileRobot Navigation in Unknown Environmentsrdquo IEEE Transac-tions on Systems Man and Cybernetics Part C (Applications andReviews) vol 37 no 4 pp 610ndash621 2007

[246] C-F Juang and Y-W Tsao ldquoA type-2 self-organizing neuralfuzzy system and its FPGA implementationrdquo IEEE Transactionson Systems Man and Cybernetics Part B Cybernetics vol 38no 6 pp 1537ndash1548 2008

[247] S Dutta ldquoObstacle avoidance of mobile robot using PSO-basedneuro fuzzy techniquerdquo vol 2 pp 301ndash304 2010

[248] A Garcia-Cerezo A Mandow and M Lopez-Baldan ldquoFuzzymodelling operator navigation behaviorsrdquo in Proceedings ofthe 6th International Fuzzy Systems Conference pp 1339ndash1345Barcelona Spain

[249] M Maeda Y Maeda and S Murakami ldquoFuzzy drive control ofan autonomous mobile robotrdquo Fuzzy Sets and Systems vol 39no 2 pp 195ndash204 1991

[250] C-S Chiu and K-Y Lian ldquoHybrid fuzzy model-based controlof nonholonomic systems A unified viewpointrdquo IEEE Transac-tions on Fuzzy Systems vol 16 no 1 pp 85ndash96 2008

[251] C-L Hwang and L-J Chang ldquoInternet-based smart-spacenavigation of a car-like wheeled robot using fuzzy-neuraladaptive controlrdquo IEEE Transactions on Fuzzy Systems vol 16no 5 pp 1271ndash1284 2008

[252] P Rusu E M Petriu T E Whalen A Cornell and H J WSpoelder ldquoBehavior-based neuro-fuzzy controller for mobilerobot navigationrdquo IEEE Transactions on Instrumentation andMeasurement vol 52 no 4 pp 1335ndash1340 2003

[253] A Chatterjee K Pulasinghe K Watanabe and K IzumildquoA particle-swarm-optimized fuzzy-neural network for voice-controlled robot systemsrdquo IEEE Transactions on IndustrialElectronics vol 52 no 6 pp 1478ndash1489 2005

[254] C-F Juang and Y-C Chang ldquoEvolutionary-group-basedparticle-swarm-optimized fuzzy controller with application tomobile-robot navigation in unknown environmentsrdquo IEEETransactions on Fuzzy Systems vol 19 no 2 pp 379ndash392 2011

[255] K W Schmidt and Y S Boutalis ldquoFuzzy discrete event sys-tems for multiobjective control Framework and application tomobile robot navigationrdquo IEEE Transactions on Fuzzy Systemsvol 20 no 5 pp 910ndash922 2012

[256] S Nurmaini S Zaiton and D Norhayati ldquoAn embedded inter-val type-2 neuro-fuzzy controller for mobile robot navigationrdquoin Proceedings of the 2009 IEEE International Conference onSystems Man and Cybernetics SMC 2009 pp 4315ndash4321 USAOctober 2009

[257] S Nurmaini and S Z Mohd Hashim ldquoMotion planning inunknown environment using an interval fuzzy type-2 andneural network classifierrdquo in Proceedings of the 2009 IEEEInternational Conference on Computational Intelligence forMea-surement Systems and Applications CIMSA 2009 pp 50ndash55China May 2009

[258] A AbuBaker ldquoA novel mobile robot navigation system usingneuro-fuzzy rule-based optimization techniquerdquo Research Jour-nal of Applied Sciences Engineering amp Technology vol 4 no 15pp 2577ndash2583 2012

[259] S Kundu R Parhi and B B V L Deepak ldquoFuzzy-Neuro basedNavigational Strategy for Mobile Robotrdquo vol 3 p 1 2012

[260] M A Jeffril and N Sariff ldquoThe integration of fuzzy logic andartificial neural network methods for mobile robot obstacleavoidance in a static environmentrdquo in Proceedings of the 2013IEEE 3rd International Conference on System Engineering andTechnology ICSET 2013 pp 325ndash330 Malaysia August 2013

[261] C Chen and P Richardson ldquoMobile robot obstacle avoid-ance using short memory A dynamic recurrent neuro-fuzzyapproachrdquo Transactions of the Institute of Measurement andControl vol 34 no 2-3 pp 148ndash164 2012

[262] M Algabri H Mathkour and H Ramdane ldquoMobile robotnavigation and obstacle-avoidance using ANFIS in unknownenvironmentrdquo International Journal of Computer Applicationsvol 91 no 14 pp 36ndash41 2014

[263] Z Laouici M A Mami and M F Khelfi ldquoHybrid Method forthe Navigation of Mobile Robot Using Fuzzy Logic and SpikingNeural Networksrdquo International Journal of Intelligent Systemsand Applications vol 6 no 12 pp 1ndash9 2014

[264] S M Kumar D R Parhi and P J Kumar ldquoANFIS approachfor navigation of mobile robotsrdquo in Proceedings of the ARTCom2009 - International Conference on Advances in Recent Tech-nologies in Communication and Computing pp 727ndash731 IndiaOctober 2009

[265] A Pandey ldquoMultiple Mobile Robots Navigation and ObstacleAvoidance Using Minimum Rule Based ANFIS Network Con-troller in the Cluttered Environmentrdquo International Journal ofAdvanced Robotics and Automation vol 1 no 1 pp 1ndash11 2016

[266] R Zhao andH-K Lee ldquoFuzzy-based path planning formultiplemobile robots in unknown dynamic environmentrdquo Journal ofElectrical Engineering amp Technology vol 12 no 2 pp 918ndash9252017

[267] J K Pothal and D R Parhi ldquoNavigation of multiple mobilerobots in a highly clutter terrains using adaptive neuro-fuzzyinference systemrdquo Robotics and Autonomous Systems vol 72pp 48ndash58 2015

[268] R M F Alves and C R Lopes ldquoObstacle avoidance for mobilerobots A Hybrid Intelligent System based on Fuzzy Logic and


Artificial Neural Networkrdquo in Proceedings of the 2016 IEEEInternational Conference on Fuzzy Systems FUZZ-IEEE 2016pp 1038ndash1043 Canada July 2016

[269] Awatef Aouf Lotfi Boussaid and Anis Sakly ldquoTLBO-BasedAdaptive Neurofuzzy Controller for Mobile Robot Navigationin a Strange Environmentrdquo Computational Intelligence andNeuroscience vol 2018 pp 1ndash8 2018

[270] R Braunstingl P Sanz and J M Ezkerra ldquoFuzzy logic wallfollowing of a Mobile Robot based on the Concept of GeneralPerceptionrdquo in Proceedings of the in 7th Int Conf on AdvancedRobotics Sant Feliu De Guixols Spain 1995

[271] C-H Hsu and C-F Juang ldquoEvolutionary robot wall-followingcontrol using type-2 fuzzy controller with species-DE-activatedcontinuous ACOrdquo IEEE Transactions on Fuzzy Systems vol 21no 1 pp 100ndash112 2013

[272] N Baklouti R John and A M Alimi ldquoInterval Type-2 FuzzyLogic Control of Mobile Robotsrdquo Journal of Intelligent LearningSystems and Applications vol 04 no 04 pp 291ndash302 2012

[273] K Al-Mutib M Faisal M Alsulaiman F Abdessemed HRamdane and M Bencherif ldquoObstacle avoidance using wall-following strategy for indoor mobile robotsrdquo in Proceedingsof the 2nd IEEE International Symposium on Robotics andManufacturing Automation ROMA 2016 Malaysia September2016

[274] M Hank and M Haddad ldquoA hybrid approach for autonomousnavigation of mobile robots in partially-known environmentsrdquoRobotics and Autonomous Systems vol 86 pp 113ndash127 2016

[275] T Takagi and M Sugeno ldquoFuzzy identification of systems andits applications tomodeling and controlrdquo IEEE Transactions onSystemsMan and Cybernetics Part B Cybernetics vol SMC-15no 2 pp 116ndash132 1985

[276] Y Zhou and M Er ldquoSelf-Learning in Obstacle Avoidance of aMobile Robot via Dynamic Self-Generated Fuzzy Q-Learningrdquoin Proceedings of the 2006 International Conference on Compu-tational Inteligence for Modelling Control and Automation andInternational Conference on Intelligent Agents Web Technologiesand International Commerce (CIMCArsquo06) pp 116-116 SydneyAustralia November 2006

[277] L Jouffe ldquoFuzzy inference system learning by reinforcementmethodsrdquo IEEE Transactions on Systems Man and CyberneticsPart C Applications and Reviews vol 28 no 3 pp 338ndash3551998

[278] C-F Juang ldquoCombination of online clustering and Q-Valuebased GA for reinforcement fuzzy system designrdquo IEEE Trans-actions on Fuzzy Systems vol 13 no 3 pp 289ndash302 2005

[279] C-F Juang and C-H Hsu ldquoReinforcement ant optimizedfuzzy controller for mobile-robot wall-following controlrdquo IEEETransactions on Industrial Electronics vol 56 no 10 pp 3931ndash3940 2009

[280] Jong-Wook Park Hwan-Joo Kwak Young-Chang Kang and DW Kim ldquoAdvanced Fuzzy Potential Field Method for MobileRobot Obstacle Avoidancerdquo Computational Intelligence andNeuroscience vol 2016 Article ID 6047906 pp 1ndash13 2016

[281] C-L Lee C-J Lin and H-Y Lin ldquoSmart robot wall-followingcontrol using a sonar behavior-based fuzzy controller inunknown environmentsrdquo Smart Science vol 5 no 3 pp 160ndash166 2017

[282] Rami Al-Jarrah Mohammad Al-Jarrah and Hubert RothldquoA Novel Edge Detection Algorithm for Mobile Robot PathPlanningrdquo Journal of Robotics vol 2018 pp 1ndash12 2018

[283] C Purcaru R Precup D Iercan L Fedorovici and R DavidldquoHybrid PSO-GSA robot path planning algorithm in static

environments with danger zonesrdquo in Proceedings of the 201317th International Conference on System Theory Control andComputing (ICSTCC) pp 434ndash439 Sinaia Romania October2013

[284] P Fabiani H-H Gonzalez-Banos J-C Latombe and D LinldquoTracking an unpredictable target among occluding obstaclesunder localization uncertaintiesrdquo Robotics and AutonomousSystems vol 38 no 1 pp 31ndash48 2002

[285] P Marin-Plaza A Hussein D Martin and A de la EscaleraldquoGlobal and Local Path Planning Study in a ROS-BasedResearch Platform for Autonomous Vehiclesrdquo Journal ofAdvanced Transportation vol 2018 pp 1ndash10 2018

[286] A Cherubini and F Chaumette ldquoA redundancy-based approachfor obstacle avoidance in mobile robot navigationrdquo in Proceed-ings of the 2010 IEEERSJ International Conference on IntelligentRobots and Systems (IROS 2010) pp 5700ndash5705 Taipei October2010

[287] A Cherubini and F Chaumette ldquoVisual navigation with obsta-cle avoidancerdquo in Proceedings of the 2011 IEEERSJ InternationalConference on Intelligent Robots and Systems (IROS 2011) pp1593ndash1598 San Francisco CA September 2011

[288] S Segvic A Remazeilles A Diosi and F Chaumette ldquoAmapping and localization framework for scalable appearance-based navigationrdquo Computer Vision and Image Understandingvol 113 no 2 pp 172ndash187 2009

[289] M Cherubini G Colafrancesco L Oriolo L Freda and FChaumette Comparing appearance-based controllers for non-holonomic navigation from a visual memory ICRAWorkshop onsafe navigation in open and dynamic environments applicationto autonomous vehicles ICRA Workshop on safe navigation inopen and dynamic environments application to autonomousvehicles 2009

[290] C Einhorn H J Schr Bohme H M Gross and H J SchroterldquoAHybrid Kalman Filter Based Algorithm for Real-Time VisualObstacle Detectionrdquo in Proceedings of the in Proceedings of the3rd European Conference on Mobile Robots (ECMR) Freiburg2007

[291] M A Moreno-Armendariz and H Calvo ldquoVisual SLAMand Obstacle Avoidance in Real Time for Mobile RobotsNavigationrdquo in Proceedings of the 2014 International Confer-ence on Mechatronics Electronics and Automotive Engineering(ICMEAE) pp 44ndash49 Cuernavaca Mexico November 2014

[292] Y Zhu R Mottaghi E Kolve et al ldquoTarget-driven visualnavigation in indoor scenes using deep reinforcement learningrdquoin Proceedings of the 2017 IEEE International Conference onRobotics and Automation ICRA 2017 pp 3357ndash3364 SingaporeJune 2017

[293] J Savage S Munoz M Matamoros and R Osorio ldquoObsta-cle Avoidance Behaviors for Mobile Robots Using GeneticAlgorithms and Recurrent Neural Networksrdquo IFAC ProceedingsVolumes vol 46 no 24 pp 141ndash146 2013

[294] A M Zaki O Arafa and S I Amer ldquoMicrocontroller-basedmobile robot positioning and obstacle avoidancerdquo Journal ofElectrical Systems and Information Technology vol 1 no 1 pp58ndash71 2014

[295] R M Santiago A L De Ocampo A T Ubando A A Bandalaand E P Dadios ldquoPath planning for mobile robots usinggenetic algorithm and probabilistic roadmaprdquo in Proceedingsof the 2017 IEEE 9th International Conference on HumanoidNanotechnology Information Technology Communication andControl Environment and Management (HNICEM) pp 1ndash5Manila December 2017


[296] Imen Hassani Imen Maalej and Chokri Rekik ldquoRobot PathPlanning with Avoiding Obstacles in Known EnvironmentUsing Free Segments and Turning Points Algorithmrdquo Mathe-matical Problems in Engineering vol 2018 pp 1ndash13 2018

[297] C H Do and H Lin ldquoDifferential evolution for optimizingmotion planning of mobile robotrdquo in Proceedings of the 2017IEEESICE International Symposium on System Integration (SII)pp 399ndash404 Taipei December 2017

[298] T Mercy E Hostens and G Pipeleers ldquoOnline motionplanning for autonomous vehicles in vast environmentsrdquo inProceedings of the 2018 15th InternationalWorkshop onAdvancedMotion Control (AMC) pp 114ndash119 Tokyo March 2018

[299] Y Li R Cui Z Li and D Xu ldquoNeural Network Approxima-tion Based Near-Optimal Motion Planning With KinodynamicConstraints Using RRTrdquo IEEE Transactions on Industrial Elec-tronics vol 65 no 11 pp 8718ndash8729 2018

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018


Active and Passive Electronic Components

VLSI Design



Shock and Vibration


Civil EngineeringAdvances in

Acoustics and VibrationAdvances in



Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of



Journal ofEngineeringVolume 2018

SensorsJournal of



RotatingMachinery


Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018


Chemical EngineeringInternational Journal of Antennas and

Propagation




Navigation and Observation


Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom


goal irrespective of the efficiency Optimality is concernedwith determining an optimized feasible plan to reach goalwith efficient performance Comparing the two achieving afeasible path planning was described to be a problem butthat of optimal path planning is more problematic Pathplanning and obstacle avoidance methods could be classifiedinto static or dynamic based on the environment and asglobal or local based on the path planning algorithm [12]Environment composed of stationary objects is described asstatic while thosewithmoving objects are termed as dynamicWhen the path planning occurs in a static environmentand all the information of the environment required bythe robot to perform path planning and obstacle avoidanceis available it is described as global path planning Localpath planning on the other hand occurs when the robotis in motion and it reacts to changes in the environmentto decide its movement [1 10] Path planning can also bedescribed as either online or off-line and reactive or mapbased [13] Off-line techniques compute the path beforenavigation while online techniques perform the computa-tions and take decision incrementally during the navigationprocess Data derived from cameras and sensors includingultrasonic sensors infrared sensors and light sensors areinterpreted in different ways by algorithms to embark on safepath planning Control of navigation and obstacle avoidanceformobile robots can be described as a reactive approach [14ndash16] because the robot is expected to react against its changingenvironment by reacting immediately to new informationreceived from sensors Control of path planning is howeverdescribed as deliberative approach because the robot isexpected to provide the exact planning to achieve a goalrelying on the geometric model of the environment and theappropriate theory The most common control method ofavoiding obstacles in cluttered environment by mobile robotis the reactive local navigation schemes where the robotrsquosaction depends on sensor information

Ongoing research in mobile robotics is geared towardsbuilding autonomous and intelligent robots that performrobust motion planning and navigation in dynamic envi-ronments [17] There are records of considerable number ofresearch that aimed at addressing path planning and obstacleavoidance problem using diverse approaches and algorithms[17ndash29] Despite the numerous efforts to address the problemof safe path planning of mobile robots there still exists achallenge in achieving real safe and optimal path planningto achieve the use of intelligent autonomous vehicles [5 30ndash35] These outstanding challenges motivated this research toconduct an overview of the popular methods and strategiesused by researchers to address the problem and to identifythe strengths and challenges of these approaches and con-sider the way forward to achieving intelligent autonomousvehicles

Mobile robot path planning methods could be catego-rized in different ways In this paper we classified theminto nature-inspired computation conventional and hybridmethods Methods and strategies which have something todo with the imitation of phenomena of nature are describedas nature-inspired computation method Those that are notrelated to imitation of nature phenomena are classified under

conventional method Methods that combine two or morestrategies are described as hybrid methods in this paper

The remainder of this paper is organized as follows dis-cussion of conventional mobile robot path planningmethodsis given in Section 2 The strengths and challenges of theconventional methods are given in Section 3 Discussion ofnature-inspired mobile robot path planning methods withtheir strengths and challenges is covered in Sections 4 and 5In Sections 6 and 7 hybrid methods with their strengths andchallenges are presented Conclusion and the way forward formobile robot path planning and obstacle avoidance are givenin Section 8

2 Conventional Path Planning Methods

The conventional path planning methods (CPPM) are thetraditional methods used over the years by researchers formobile robot path planning Most of these methods rely ondistance information from objects to the robots force ofattraction and repulsion statistical methods clustering orgraphical map calculations to determine the path planning ofthe robotThey are comparatively computationally expensiveNotable among them are artificial potential field (APF)distance control algorithm bumper event approach wall-following sliding mode control (SMC) dijkstra Alowast DlowastStereo block matching voronoi diagram (VD) simultaneouslocalization and mapping (SLAM) vector field histogram(VFH) rapidly exploring random tree (RRT) curvaturevelocity lane curvature dynamic window and tangent graphAccounts of the use of some of these approaches are given inthe next subsections

21 Artificial Potential Field Path Planning Method One ofthe popular methods used by researchers over the years isAPF APF is a mathematical method which causes a robotto be attracted to the goal but repelled by obstacles in theenvironment [6 34 36 37] There are notable research workscarried out using APF with sensors to plan the path of mobilerobots to provide autonomous navigation taking obstacleavoidance into consideration [12 37ndash47]The idea of applyingAPF to path planning originated from Khatib [34] APFapproach was modified by some researchers to make it moreefficient for path planning and obstacle avoidance [41 4348 49] The algorithm of APF comprises the attractive forcerepulsive potential force and the total potential force APFalgorithmusingGaussian function as in [50] can be expressedas shown in the following

119860119865 (119901) = 119891119886119905119905 [1 minus 119864119883119875 (minus119888119886119905119905 lowast 1198892119892)] (1)

119877119865 (119901)

= sum

119891119903119890119901119894 [119864119883119875 (minus119888119903119890119901 lowast 1198892119900119887119904119894)] 119889119900119887119904119894 (119901) le 11988900 119890119897119904119890

(2)

119865119905119900119905119886119897 (119901) = 119860119865 (119901) + 119877119865 (119901) (3)

where AF(p) is total attraction force fatt is maximum value ofattraction force at any instance catt is attractive constant dg is





































































(4)

119910119894 (119896 + 1) = 119910119894 (119896) + 119881119894 (119896 + 1) (5)
























(a) (b)


6 Hybrid Methods





































































with real robot







Simulation




Simulation
































Simulation andreal

Experiments








References





























































































































































































































































































































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia






































































(4)

119910119894 (119896 + 1) = 119910119894 (119896) + 119881119894 (119896 + 1) (5)
























(a) (b)


6 Hybrid Methods





































































with real robot







Simulation




Simulation
































Simulation andreal

Experiments








References





























































































































































































































































































































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia















(a) (b)


6 Hybrid Methods





































































with real robot







Simulation




Simulation
































Simulation andreal

Experiments








References





























































































































































































































































































































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia






























































with real robot







Simulation




Simulation
































Simulation andreal

Experiments








References





























































































































































































































































































































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia




























Simulation andreal

Experiments








References





























































































































































































































































































































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia







References





























































































































































































































































































































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia






































































































































































































































































































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia


an overview of nature-inspired, conventional, and hybrid...

Documents