360 degree conops · executive summary this report details the design, deployment, and evaluation...
TRANSCRIPT
Implementation and Evaluation of the
360 Degree Radar Information System
Prepared By:
Daniel M. Nelson, AICP
AECOM
July 2016
Published By:
Minnesota Department of Transportation
1500 West County Road B2
Roseville, MN 55113
This report represents the results of research conducted by the authors and does not necessarily represent the
views or policies of the Minnesota Department of Transportation or AECOM or RhiZone. This report does not
contain a standard or specified technique.
The authors, the Minnesota Department of Transportation, and AECOM do not endorse products or
manufacturers. Any trade or manufacturers’ names that may appear herein do so solely because they are
considered essential to this report.
Table of Contents
Executive Summary .................................................................................................................. 1 Chapter 1 Introduction ......................................................................................................... 1
1.1. Project Purpose ........................................................................................................... 1 1.2. Project Background .................................................................................................... 1 1.3. Operational Needs ...................................................................................................... 2
1.4. Project Description ..................................................................................................... 2 1.5. Report Organization ................................................................................................... 3
Chapter 2 Systems Engineering for ITS Process ................................................................. 4 2.1. MnDOT Statewide Regional ITS Architecture and ITS Development Objectives ... 5 2.2. Systems Engineering Documents ............................................................................... 7
Chapter 3 360 Radar Information System Equipment......................................................... 8 3.1. 360 Degree Radar and CCTV Camera Equipment .................................................... 8
3.2. FCC License for 360 Radar Unit.............................................................................. 10 3.3. Central Office Equipment ........................................................................................ 10 3.4. Central Software Calibration .................................................................................... 12 3.5. Overall System Installation and Adjustment Timeline ............................................ 14
Chapter 4 360 Degree Radar System Evaluation .............................................................. 16 4.1. System Validation Process ....................................................................................... 16 4.2. System Validation Results ....................................................................................... 17
4.3. System Evaluation Process....................................................................................... 27 4.4. System Evaluation Results ....................................................................................... 28
Chapter 5 Conclusions and Recommendations ................................................................. 57 5.1. Conclusions .............................................................................................................. 57 5.2. Considerations for Future Installation ...................................................................... 59
5.3. Recommendations for Future Installation ................................................................ 63
Appendix A – System Component Data Sheets ..................................................................... 65
Acknowledgements
The authors of this report wish to acknowledge the very helpful efforts of several MnDOT
staff at the Regional Transportation Management Center (RTMC) that helped make this
Innovative Idea project a successful project, including Project Manager Rashmi Brewer,
RTMC Operations staff Brian Kary, Terry Haukom, Ralph Adair, and Jesse Larson, as well
as technical support received from MN.IT technician Tim Johnson.
Figures in Document
Figure 1-1 – General Problems and Operational Needs ........................................................... 2 Figure 1-2 – 360 Degree Radar Information System Overview ............................................... 3 Figure 2-1 – Systems Engineering V-Diagram ......................................................................... 4 Figure 3-1 – 360 Degree Radar and CCTV Cameras ............................................................... 9 Figure 3-2 – Diagram of 360 Radar Equipment and Locations on Pole ................................... 9
Figure 3-3 –Central Office Components in RTMC Server Room .......................................... 10 Figure 3-4 – Witness Software Interface of Corridor Sections and Vehicles ......................... 11 (July through September 2015) ............................................................................................... 11 Figure 3-5 – Witness Software Interface of Corridor Sections............................................... 11 (December 2015 through March 2016) ................................................................................... 11
Figure 3-6 – NVR Interface with Camera Views of Corridor ................................................ 12 Figure 3-7 – Current Settings with Central Software ............................................................. 13
Figure 4-1 – MnDOT Computer Aided Dispatch Summary of Dec. 17th Event .................... 18 Figure 4-2 – MnDOT Camera Record of Dec. 17th Event for Confirmation with ................. 18 360 Degree Radar System ....................................................................................................... 18 Figure 4-3 – 360 Degree Radar System Record of Dec. 17th Event ....................................... 19
Figure 4-4 – Interface of NVR with Video Record Dec. 17th Event ..................................... 20 Figure 4-5 – Navtech and Wavetronix WB and EB Speed Comparisons on Nov. 20th .......... 30 Figure 4-6 – Navtech and Wavetronix WB and EB Speed Comparisons on Nov. 20th .......... 33
By Lane for all Time Periods .................................................................................................. 33 Figure 4-7 – Navtech, Wavetronix, and Manual WB and EB Count Comparisons on Nov.
19th .......................................................................................................................................... 34 Figure 4-8 – Navtech and Wavetronix WB and EB Count Comparisons on Nov. 19th By
Lane for all Time Periods ....................................................................................................... 38
Figure 4-9 – Navtech and Wavetronix WB and EB Speed Comparisons on Jan. 17th, 2016 . 39
Figure 4-10 – Navtech and Wavetronix WB and EB Speed Comparisons on Jan. 17th By
Lane for all Time Periods ....................................................................................................... 42 Figure 4-11 – Navtech and Wavetronix WB and EB Count Comparisons on Jan. 17th ......... 43
Figure 4-12 – Navtech and Wavetronix WB and EB Count Comparisons on Jan. 17th By
Lane for all Time Periods ....................................................................................................... 47
Figure 4-13 – Navtech and Wavetronix WB and EB Speed Comparisons on Feb. 2nd, 2016 48 Figure 4-14 – Navtech and Wavetronix WB and EB Speed Comparisons on Feb. 2nd By
Lane for all Time Periods ....................................................................................................... 51 Figure 4-15 – Navtech and Wavetronix WB and EB Count Comparisons on Feb. 2nd .......... 52
Figure 4-16 – Navtech and Wavetronix WB and EB Count Comparisons on Feb. 2nd By
Lane for all Time Periods ....................................................................................................... 56
Tables in Document
Table 2-1 – Rule 940 Requirements ......................................................................................... 5 Table 2-2 – Minnesota ITS Development Objectives Related to 360 Degree Radar
Information System ................................................................................................................... 6
Table 2-2 – Systems Engineering Documents .......................................................................... 7 Table 3-1 – System Component Model Numbers and Detail ................................................... 8 Table 3-2 – High-Level System Installation and Adjustment Timeline ................................. 14 Table 4-1 – Validation Plan Measures of Effectiveness ......................................................... 16 Table 4-2 – Summary of Events Logged by 360 Degree Radar and MnDOT CAD System
(August – September 2015) .................................................................................................... 22 Table 4-3 – Summary of Events Logged by 360 Degree Radar and MnDOT CAD System
(February - March 2016) ......................................................................................................... 23 Table 4-4 – Summary of Events Logged by 360 Degree Radar and MnDOT CAD System . 24
Table 4-5 – Summary of False Alarms Logged June 20th to June 26th ................................... 25 Table 4-6 – Summary of False Alarms Logged March 2nd to March 9th ................................ 26
Table 4-7 – Summary of Person and Reverse Vehicle Alarms Verified June 20th to June 26th
................................................................................................................................................. 27
Table 4-8 – Dates of Vehicle Speed and Count Comparisons with Wavetronix Data ........... 27 Table 5-1 – Summary of Conclusions on 360 Degree Radar Information System ................ 57 Table 5-2 – Summary of Key Findings from Evaluation of Incident Detection Capabilities
the 360 Degree Radar Information System ............................................................................. 58 Table 5-3 – Summary of Operational Needs Addressed by the .............................................. 58
360 Degree Radar Information System ................................................................................... 58
Executive Summary
This report details the design, deployment, and evaluation of the 360 Degree Radar Information
System that has been installed along the side of Interstate 94 in Minneapolis between the 3rd Ave.
S. and MN-65 overpasses. The purpose of the project is to evaluate the system for MnDOT and
determine whether the system could reliably be used by operators at the MnDOT Regional
Transportation Management Center (RTMC) for automated incident detection and dispatch of
emergency response personnel.
MnDOT selected AECOM and RhiZone under the Innovative Idea Program in June 2014 to plan
for and install the 360 Degree Radar Information System. AECOM utilized the Systems
Engineering process to guide the project through the concept, design, installation, and evaluation
stages. RhiZone coordinated with the system manufacturer – Navtech Radar Surveillance
(Navtech) – to procure, configure and install the system which consisted of a single 360 Degree
Radar unit as field equipment, as well as a central software package that operated on a server at
the MnDOT RTMC in Roseville, MN.
The system was initially installed on June 2nd, 2015 and initial configurations to field equipment
along with central software were then completed by June 19th, 2015 to allow for the operational
test to begin. Subsequent configuration adjustments were made in the following months to
improve the accuracy of incident detection and measurements of vehicle counts and speeds.
The purpose of this project was to evaluate the system operation for MnDOT after system
installation and determine whether the system could reliably be used by operators at the MnDOT
RTMC for automated incident detection and dispatch of emergency response personnel.
AECOM performed the evaluation between July 2015 and March 2016 and CCTV cameras were
installed for the purpose of recording incidents detected by the 360 Degree Radar Information
System to measure the amount of false alarms generated by the system.
Three key findings of the evaluation related to automated incident detection are summarized
below:
1. The 360 Degree Radar Information System would have provided an average of 10.5
minutes of advance notification to RTMC operators of incidents in the project area (based
on 11 total events logged in the project for comparison with the MnDOT CAD system).
This shorter notification time would help accomplish a MNDOT safety objective of
advanced notification of incidents.
2. The false alarm rate of incident notification decreased from 0.71 at the beginning of the
project to 0.28 near the end of the project, resulting in a positive sign that the 360 Degree
Radar Information System can reduce its false alarm rate over the course of its operation.
3. Between August and September 2015, the 360 Degree Radar Information System
detected 11 events that were not logged by RTMC operators as traffic incidents. A
notification to RTMC operators of these events would also greatly help accomplish the
higher level safety objectives for MnDOT of reducing injuries, fatalities, and the potential
for secondary crashes on the corridor.
AECOM also performed an evaluation between July 2015 and March 2016 of the accuracy of
vehicle counts and speeds as measured by the 360 Degree Radar Information System. While the
ability to collect accurate vehicle counts and speeds was not the primary objective of the project,
Navtech Radar was able to upgrade the central software to allow for the system to
simultaneously perform this task alongside the incident detection features of the overall system.
Two key findings of the evaluation related to vehicle counts and speeds are summarized below:
1. The vehicle speeds recorded by the system were within 10% of the vehicle speeds
measured by the Wavetronix unit for all lanes of traffic measured in November 2015.
2. The vehicle counts recorded by the system were within 10% of the vehicle speeds
measured by the Wavetronix unit at the field site for the right lane of traffic in each
direction in November 2015. System vehicle counts were consistently lower than the
Wavetronix measured data by greater than 10% for the middle and fast lanes of traffic.
3. Software configurations performed in January 2015 impacted the ability to accurately
compare vehicle counts and speeds measured by the system in January and February
2016 with Wavetronix data.
4. The accuracy of the counts and speeds in the westbound lanes was about 15-20% more
accurate when compared with Wavetronix data than count and speed measurements of
the eastbound lanes, and this was expected due to the greater distance of those eastbound
lanes from the radar unit.
After the evaluation of the incident detection capabilities and the accuracy of the vehicle counts /
speeds were performed, MnDOT allowed for a limited trial period of the system within the
MnDOT RTMC for two weeks in late May and early June. The system was monitored by a
MnDOT traffic operations supervisor who noted that many stopped vehicle alarms were
generated during periods of heavy traffic congestion on the corridor. As a result, the value of
alarms for incident detection was limited given the frequency of traffic congestion in the area.
As a lesson learned for this traffic environment, Navtech proposed a modification to the software
that could limit the number of alarms generated in heavy traffic congestion.
The biggest challenge encountered on the project by the AECOM team was properly configuring
both the radar unit in the field and the central software installed at the MnDOT RTMC to
accurately perform the simultaneous tasks of incident detection and measurements of vehicle
counts and speeds on the corridor. The proper tilt of the radar toward the roadway was initially
established but then modified slightly in July in an effort to improve the accuracy of vehicle
detection in the project area. This configuration of the radar unit slightly limited the radar’s
capability to detect stopped vehicles on the corridor.
In addition, the presence of the MN-65 bridge overpass and the 3rd Avenue South bridge
overpass presented an environment that Navtech had not previously tested with the 360 Degree
Radar Information System. It was also learned through the course of the project that heavy
traffic congestion in the project area caused the central software to ignore multiple alarms of
stopped vehicles at the same time. This had the impact of causing the software to ignore stopped
vehicle events reported by the radar unit that were actual incidents logged by the RTMC
operators. Navtech Radar learned of this issue through this project and is actively developing a
software upgrade to address this issue in future deployments.
Configurations to the central software were also performed in October in an effort to improve the
accuracy of vehicle counts and speeds on the corridor, while also simultaneously performing the
main task of incident detection. This configuration required software upgrades from Navtech
Radar to perform these simultaneous tasks of incident detection and vehicle counts, which
impacted the ability to perform an evaluation of the system operations during the period of the
software updates.
Through the overall project, the following advantages and limitations were observed:
System Advantages System Limitations
1. Speed of detection – Ability to
quickly detect objects and track the
progress of movement through
project area
1. Occlusion – Radar requires line-of-sight with
object for detection. Large semi-trucks can
block smaller vehicles from being detected
2. Non-intrusive detection – Detection
of objects does not require in-
pavement loops or other hardware
for operations
2. Configuration – Software configuration of
detection zones and rules for detecting objects
requires careful monitoring in the initial period
of the system installation. Changes to system
settings also require time to confirm that
accurate detection is being performed.
3. Maintenance – Low amount of
hardware maintenance;
recommended to change an internal
belt once every three years.
3. System Cost – Cost of the field equipment and
software for the one location in this project was
approx. $50,000, which could be used for other
safety countermeasures such as more Service
Patrol vehicles. Navtech has been made aware
of the cost concerns and has noted future
deployments would need to have reduced costs,
as well as a small monthly fee to make the
radar cost effective at detecting incidents.
Given the overall performance of the 360 Degree Radar Information system on the project, the
following next steps are envisioned for future system installations:
1. Operation at Signalized Intersections – The system could be used to monitor signalized
intersections given its noted accuracy in detecting specific events and vehicle presence.
The system could replace in-pavement loops or other technology currently installed to
actuate signal operations. In terms of safety, the system’s ability to detect pedestrians
could help reduce pedestrian-vehicle crashes and injuries at those intersections.
2. ICWS Integration – The system could also be integrated with Intersection Conflict
Warning System (ICWS) installed by MnDOT at rural locations. The proven ability of
the system to detect vehicles traveling at high speeds from similar distances as the current
ICWS installations could prove to be more cost effective than previous installations while
maintaining or improving the safety of travel at uncontrolled intersections with a previous
history of vehicle incidents
1
Chapter 1 Introduction
This final report details the design, deployment, and evaluation of the 360 Degree Radar
Information System that has been installed along the side of Interstate 94 in Minneapolis
between the 3rd Ave. S. and MN-65 overpasses.
The 360 Degree Radar Information System is designed to generate automated alerts of traffic
incidents to Traffic Management Center (TMC) operators, who could then verify the incident
through use of CCTV cameras and, in turn, dispatch emergency response personnel in a more
timely and efficient manner than waiting for a 911 emergency call from those involved in the
incident.
1.1. Project Purpose
The purpose of this project is to install and evaluate the operation of the 360 Degree Radar
Information System for MnDOT over a 6-month period and determine whether the system could
reliably be used by operators at the MnDOT Regional Transportation Management Center
(RTMC) for timely incident detection and dispatch of emergency response personnel.
The area of Interstate 94 where this system has been installed experiences very high volumes of
vehicular traffic and the highest amount of vehicle crashes and incidents in the state of
Minnesota, given its convergence with Interstate 35W and proximity to the Lowry Hill tunnel.
The ability to quickly dispatch emergency response personnel based on timely automated alerts
from the 360 Radar Information System could reduce the number of vehicular fatalities, injuries,
and property damage that may result from the primary crash and potential secondary crashes as
well.
1.2. Project Background
The 360 Degree Radar Information System was submitted as a proposal in response to a MnDOT
solicitation under the Innovative Idea program. A detailed proposal was submitted by AECOM
(and sub-contractor RhiZone) to MnDOT on June 3rd, 2014 as part of Stage II of the program
that illustrated the plan to install the 360 Degree Radar Information System. The use of the
Systems Engineering (SE) process for Intelligent Transportation Systems (ITS) was also
proposed to guide the project through the concept, design, installation, and evaluation stages of
the project.
An optional application of installing a second 360 Radar Information System at a signalized
intersection was also proposed, however, only as an option for MnDOT to consider. The current
system design does not anticipate installing the second system at an intersection for traffic
monitoring, although, this could be an option for MnDOT to consider in future testing of the
system.
The use of 360 Degree Radar Systems have been tested and are operational in Europe, the
Middle East, and Australia. The manufacturer of the equipment and software is known as
Navtech, and this project is the first demonstration of the radar unit in the United States. An
2
experimental license from the Federal Communications Commission (FCC) was obtained locally
by RhiZone prior to system installation.
1.3. Operational Needs
Current methods of traffic and vehicle detection have various limitations that are summarized in
the figure below. The operational needs for the 360 Degree Radar Information System are also
summarized with respect to each of the general problems noted in Figure 1-1.
Figure 1-1 – General Problems and Operational Needs
General Problems Operational Needs
1. Limited zone coverage – Current methods of
loop detection are spot-based with loops installed
at intervals in lanes of traffic.
1. Deploy a traffic detection system
that can monitor multiple lanes of
traffic from one single location.
2. Understanding of latency in incident detection –
It is difficult to understand the latency in time
between when traffic incidents occur on a corridor
and when a formal incident notification is received
by emergency responders
2. Gather incident detection
information quickly and accurately in
the project area and provide
information to operational staff for
emergency response
3. Inability to measure driver behavior; individual
vehicle movements – Loop detection and other
non-intrusive detection methods measure traffic
volumes as a whole along the roadway, and do not
allow the capability for measuring driver
behaviors to potentially understand how
congestion forms on a roadway.
3. Gather traffic congestion information
quickly and accurately in the project
area
4. Inability to perform multiple real-time data
collection tasks over a wide area – Loop detection
and other non-intrusive detection methods only
allow for vehicle count measurements at those
locations.
4. Understand the amount of lane-
weaving occurring within the project
area
5. Difficult to gather performance metrics in a cost
effective manner
5. Gather performance metrics that can
be reported regarding traffic congestion
and incidents in the project area in a
readable and usable format
1.4. Project Description
Field equipment includes the 360 Degree radar unit and independent CCTV cameras to be used
for verifying incidents detected by the radar unit along the side of I-94 between the 3rd Ave. S.
and MN-65 overpasses. This equipment utilizes MnDOT fiber-optic cable to communicate data
to a System Server installed within the MnDOT RTMC Server Room. Data from the radar and
video from the CCTV cameras are recorded on a Network Video Recorder (NVR) for the
purposes of recording information communicated from the 360 Degree Radar Information
System in the field. In addition, data and video can be viewed locally within the MnDOT RTMC
Server Room via an existing MnDOT workstation in the server room, as well as through remote
3
internet access Figure 1-2 contains an illustration of the field and central office elements within
the system.
Figure 1-2 – 360 Degree Radar Information System Overview
1.5. Report Organization
This report is organized into the following Chapters:
Chapter 2: Systems Engineering for ITS Process
Chapter 3: 360 Degree Radar Information System Equipment
Chapter 4: 360 Degree Radar System Evaluation
Chapter 5: Conclusions and Recommendations
4
Chapter 2 Systems Engineering for ITS Process
This Chapter describes the Systems Engineering for ITS process that was followed in planning,
designing, installing, and testing the 360 Degree Radar Information System. This process is
highly recommended by the Federal Highway Administration (FHWA), which helps to guide
ITS implementations, such as the 360 Degree Radar Information System, by involving
stakeholders early and developing their ideas before high-level and detailed design activities are
conducted. This interdisciplinary approach assures that the ultimate design and implementation
of the system reflects this early input. The Systems Engineering “V” Diagram is provided in
Figure 2-1 below.
Figure 2-1 – Systems Engineering V-Diagram
Systems Engineering integrates all the disciplines and specialty groups into a team effort forming
a structured development process that proceeds from system concept to implementation and
operation. Systems Engineering considers both the business and the technical needs of all
customers with the goal of providing a quality product that meets the user needs.
The Systems Engineering process is used to identify a project’s needs and constraints and lay out
the activities, resources, budget, and timeline for the project. A critical part of the process is to
build consensus among the stakeholders of the project. The process is applicable at all stages of a
project, from initial system planning through final operations and maintenance of the system.
FHWA Federal Rule 940, Intelligent Transportation Systems Architecture and Standards, which
implements Section 5206 (e) of the Transportation Equity Act of the 21st Century (TEA – 21),
requires agencies implementing projects with ITS elements utilizing federal funds to develop
5
regional architectures and adopt a Systems Engineering approach for project deployments in
order to qualify for ITS grants. Table 2-1 illustrates the relationship between the various steps of
the SE process and Rule 940.
Table 2-1 – Rule 940 Requirements
Systems Engineering
Process Steps
Corresponding Rule 940 Requirements
Concept of Operations
Identification of participating agencies roles and
responsibilities
Procedures and resources necessary for operations and
management of the system
System Requirements:
High-Level and Detailed Requirements definitions
Design:
High-Level and Detailed
Identification of portions of the regional ITS architecture
being implemented
Analysis of alternative system configurations and
technology options to meet requirements
Procurement options
Identification of applicable ITS standards and testing
procedures
2.1. MnDOT Statewide Regional ITS Architecture and ITS Development Objectives
MnDOT has developed a Statewide Regional ITS Architecture that provides the base for all ITS
projects illustrated in the Systems Engineering diagram in Figure 2-1. From the MnDOT
Statewide Regional ITS Architecture in 2014, MnDOT’s ITS Development Objectives were
refined to better align with Minnesota’s ITS needs and be consistent with the National ITS
Architecture.
The goal of the Minnesota ITS Development Objectives is to enhance transportation through the
safe and efficient movement of people, goods, and information, with greater mobility, fuel
efficiency, less pollution and increased operating efficiency statewide. These ITS Development
Objectives are rooted from the overall Minnesota GO Transportation Plan that defines MnDOT’s
vision for transportation and guiding principles and outlines strategies to satisfy its vision and
mission. The 20-Year Statewide Multimodal Transportation Plan (SMTP) further clarifies these
strategies and lays out actions to implement the strategies. The Minnesota ITS Development
Objectives presented in Table 2-2, establishes the specific objectives that can be achieved
through implementing ITS countermeasures, such as the 360 Degree Radar Information System.
6
Table 2-2 – Minnesota ITS Development Objectives Related to 360 Degree Radar
Information System A. Improve the Safety of the State's Transportation System
A-1 – Reduce crash frequency
A-1-01 Reduce number of vehicle crashes
A-1-02 Reduce number of vehicle crashes per VMT
A-1-04 Reduce number of crashes due to unexpected congestion
A-1-05 Reduce number of crashes due to red-light running
A-1-08 Reduce number of crashes due to inappropriate lane departure, crossing and merging
A-1-10 Reduce number of crashes at signalized intersections
A-1-11 Reduce number of crashes at un-signalized intersections
A-1-15 Reduce number of crashes involving pedestrians and non-motorized vehicles
A-1-19 Reduce number of all secondary crashes
A-2 – Reduce fatalities and life changing injuries
A-2-01 Reduce number of roadway fatalities
A-2-02 Reduce number of roadway fatalities per VMT
A-2-04 Reduce number of fatalities due to unexpected congestion
A-2-05 Reduce number of fatalities due to red-light running
A-2-09 Reduce number of fatalities due to inappropriate lane departure, crossing and merging
A-2-11 Reduce number of fatalities at signalized intersections
A-2-12 Reduce number of fatalities at un-signalized intersections
A-2-16 Reduce number of fatalities involving pedestrians and non-motorized vehicles
A-2-22 Reduce number of roadway injuries
A-2-23 Reduce number of roadway injuries per VMT
A-2-25 Reduce number of injuries due to unexpected congestion
A-2-30 Reduce number of injuries due to inappropriate lane departure, crossing and merging
A-2-26 Reduce number of injuries due to red-light running
A-2-32 Reduce number of injuries at signalized intersections
A-2-33 Reduce number of injuries at un-signalized intersections
A-2-37 Reduce number of injuries involving pedestrians and non-motorized vehicles
B. Increase Operational Efficiency and Reliability of the Transportation System
B-1 – Reduce overall delay associated with congestion
B-1-01 Reduce the percentage of facility miles (highway, arterial, rail, etc.) experiencing recurring
congestion during the peak period
B-1-02 Reduce the percentage of Twin Cities freeway miles congested in weekday peak periods
B-1-03 Reduce the share of major intersections operating at LOS F
B-1-06 Reduce the number of hours per day that the top 20 most congested roadways experience
recurring congestion
B-1-10 Reduce hours of delay per capita
B-1-11 Reduce hours of delay per driver
B-1-16 Reduce mean time for needed responders to arrive on-scene after notification
B-1-17 Reduce mean incident clearance time per incident (Defined as the time between awareness of an
incident and the time the last responder has left the scene.)
B-1-18 Reduce mean incident clearance time for Twin Cities urban freeway incidents (Defined as the
time between awareness of an incident and the time the last responder has left the scene.)
C. Enhance Mobility, Convenience, and Comfort for Transportation System Users
C-1 – Reduce congestion and incident-related delay for travelers
C-1-09 Increase number of regional road miles covered by ITS-related assets (e.g., roadside cameras,
dynamic message signs, vehicle speed detectors) in use for incident detection / response
7
2.2. Systems Engineering Documents
The documents developed by the AECOM team for MnDOT are listed and described in Table 2-
2 below. Documents were shared with MnDOT team members and updated based on input and
feedback from team members. Final versions of the documents have been produced and shared
with MnDOT.
Table 2-2 – Systems Engineering Documents
Systems Engineering
Documents Purpose/Description Deliverables
Project Plan Describes all of the tasks that need to be performed
to accomplish the project.
Mid-December
2015
System Engineering
Management Plan
(SEMP)
Addresses project controls that are to be developed
and implemented throughout the project and
documents the technical processes to be used.
Mid-January 2015
Concept of Operations Describes how the system will operate and outlines
the roles and responsibilities of each stakeholder. Mid-January 2015
Verification and
Validation Plan
Defines the step by step procedures to conduct
verifications that components meet system
requirements. Describes how the goals and
objectives of the system (identified in the ConOps)
will be measured
Mid-January 2015
System Requirements
and Design
Describes what the system will do and how its
components will function at a high level. Also
includes the development of high-level design
documents and detailed design plans including shop
drawings and specifications of the proposed system
System
Requirements in
February 2015;
System Design in
Mid-April 2015
Acceptance Testing
and Evaluation Plans
Documents the detailed procedures to be used to
verify and validate the deployed 360 Degree Radar
Information System.
June 2015
Final Project Report
Provides overall summary of activity and the
recommendations for MnDOT on next steps that
could be taken. Includes results measured during
the Acceptance Testing and Evaluation stages of the
project, as well as results from the Validation stage
of the project.
April 2016
8
Chapter 3 360 Radar Information System Equipment
A summary of the system components and quantities to be provided by the AECOM Team and
installed is presented in Table 3-1. Refer to Appendix A of this document for datasheets on the
various system components with more detailed information.
Table 3-1 – System Component Model Numbers and Detail
System
Components Unit Type Quantity Notes
360 Degree Radar CTS350-X 1 Refer to Appendix A for datasheet with
detailed component data
CCTV Cameras Axis Q1614-E 2 Current lens offered only provides 80
degree horizontal field of view
System Server Navtech Server 1
To be provided by Navtech with radar
unit and cameras as one system. Refer to
Appendix A for more detailed data on
server.
System NVR Unit Samsung NVR 1
To be implemented by Navtech on the
System Server for video recording. Refer
to Appendix A for more information.
Recording
Trigger Module Adams 1
Implemented to trigger a video recording
of an incident detected by the system
software
System Software Witness
Software 1
To be implemented by Navtech on the
System Server with graphical user
interface for viewing of radar and
cameras.
3.1. 360 Degree Radar and CCTV Camera Equipment
The 360 Degree Radar unit and CCTV cameras are pictured in Figure 3-1 as they have been
installed in the field. The radar and cameras connect via Ethernet cables into a MnDOT operated
Cisco switch in a cabinet on the pole. Figure 3-2 also illustrates the general height of the
equipment in relation to other equipment previously installed on the pole. The optimal height for
the radar to operate in the field is 12 to 15 feet above the highest point of the lanes of travel on
the I-94 corridor. Since the highest point of the roadway is at the same level as the ground at the
base of the pole, the 360 degree radar unit was installed on a bracket just above the hinge on the
pole which has been measured at about 13 feet above ground level.
Once the radar unit on the pole is properly leveled on its mounting bracket, it requires very low
maintenance to continue to operate. The manufacturer of the radar recommends that a rubber
belt used to spin the radar device 4 times per second be replaced once every three years. This
process of replacing the belt could likely be performed in the field after removing the radar from
the bracket. The re-installation process of the radar after the belt replacement would also take
9
less time, since the upper plate of the bracket on which the radar rests would not need to be re-
leveled.
The University of Minnesota Center for Transportation Studies (CTS) has installed additional
radar equipment on the pole for research and data collection purposes. During the course of time
when the University’s equipment was installed on the pole, there were no reported issues in the
operation of the two radar units operating in close proximity given their different operating
frequencies. The University’s equipment was removed from the pole in July 2015.
CCTV Cameras
(Ports 1+2)
360 Degree
Radar Unit
(Port 3)
Figure 3-1 – 360 Degree Radar and CCTV Cameras
Figure 3-2 – Diagram of 360 Radar Equipment and Locations on Pole
10
3.2. FCC License for 360 Radar Unit
As noted earlier, the 360 Degree Radar Information System has not been tested or deployed in
the United States, and thus required approval from the FCC prior to field installation and
operation. The application was formally submitted on February 10th, 2015 and approved by the
FCC on March 10th, 2015 under the call sign WH2XQL.
The license only allows for the use of the radar at the current field location. Additional
modifications can be proposed to the license that would allow for the movement of the radar to
an additional location, such as an intersection which had been proposed as an optional work task.
The current expiration date of the license is March 1st, 2017, but could be extended through a
renewal application.
3.3. Central Office Equipment
The equipment installed at the MnDOT RTMC Server Room is displayed in Figure 3-3 below.
Field equipment communicates via Ethernet cables to a MnDOT operated Cisco switch which is
also connected to MnDOT fiber-optic cable leading to the MnDOT RTMC.
Incident Trigger Module
Network Video Recorder
Navtech Server with Software
Figure 3-3 –Central Office Components in RTMC Server Room
The Navtech Server houses the Witness Software program that receives and processes data from
the radar unit on the pole. The software program determines when alarms should be generated to
alert a TMC operator of the following types of events:
Stopped vehicles on the road for a period of 10 seconds or greater
Slow-moving vehicles on the road at 15 MPH or less
Vehicle moving in Reverse on the road
People detected walking on the road
11
These alarms can be configured so as not to set off multiple alarms for similar events in
succession, and they can also be delayed. Alarms are also segmented on the corridor as shown in
Figure 3-4. This alerts an operator to focus on a specific area of the corridor where an event is
detected and determine what type of emergency response is needed. For this project, the CCTV
cameras have the best view of sections 2 and 3 in the westbound direction of I-94 and sections 6
and 7 in the eastbound direction of I-94. The circular objects in each section represent vehicles
detected by the radar. The red circular object represents the location of the radar unit on the
corridor. It should be noted that data from these sections were collected between July and
September 2015, after which time a different set of sections were implemented following updates
to the system software to improve the accuracy of vehicle counts and speeds. The updated
sections are presented in Figure 3-5.
Figure 3-4 – Witness Software Interface of Corridor Sections and Vehicles
(July through September 2015)
Figure 3-5 – Witness Software Interface of Corridor Sections
(December 2015 through March 2016)
12
When the software detects one of the four events as noted earlier, it sends a message to the
Incident Trigger Module, which in turn, instructs the Network Video Recorder (NVR) to begin a
recording of the event that was detected. The recording captures 15 seconds of video before the
event was detected, and 30 seconds of time after the event was detected. While the NVR is
always operating to record video, it only saves videos when instructed by the trigger module and
it buffers the remaining video to conserve storage space on the NVR.
The purpose of recording videos of events detected by the radar unit and Witness software is to
verify for MnDOT that the alarms presented to RTMC operators are reliable to act upon, and that
only a very low number of “false alarms” are created. This will help to create trust in the
reliability of the software and, in turn, reduce incident notification times for incidents detected on
the corridor if the equipment were deployed over a longer term. An image of the interface
available with the NVR unit is presented in Figure 3-6. Each of the two cameras present fixed
views of the corridor with date / time stamps that can be verified against the date / time stamps of
events logged by the radar unit and software.
Figure 3-6 – NVR Interface with Camera Views of Corridor
3.4. Central Software Calibration
The Witness software that analyzes the radar data communicated from the field requires a lot of
fine calibration during the initial setup of the system as shown in the screenshot taken below in
Figure 3-7. The settings displayed in the Figure represent the most recent settings implemented
for the project. The challenge is in setting the “Track Density for Queue”, the “Slow Density for
Queue”, and the “Minimum Speed for Queue” to detect traffic incidents while trying not to
detect general traffic congestion as much.
It should be noted that Navtech Radar is currently developing on a software upgrade to better
detect traffic incidents during periods of heavy traffic congestion, which came as a direct result
from their experience on this project.
13
Westbound Traffic Eastbound Traffic
Figure 3-7 – Current Settings with Central Software
14
3.5. Overall System Installation and Adjustment Timeline
The biggest challenge encountered on the project by the AECOM team was properly configuring
both the radar unit in the field and the central software installed at the MnDOT RTMC to
accurately perform the simultaneous tasks of incident detection and measurements of vehicle
counts and speeds on the corridor. The proper tilt of the radar toward the roadway was initially
established but then modified slightly in July in an effort to improve the accuracy of vehicle
detection in the project area. This configuration slightly limited the radar’s capability to detect
stopped vehicles on the corridor. Configurations to the central software were also performed in
October in an effort to improve the accuracy of vehicle counts and speeds on the corridor. This
configuration required software upgrades from Navtech Radar to perform these simultaneous
tasks of incident detection and vehicle counts.
A summary of installation and adjustment activities is presented in Table 3-2. A key event in the
timeline of activity was the adjustments made to the central software to improve the radar’s
ability to measure vehicle counts and speeds in early October 2015. This created a level of
uncertainty in the radar’s ability to accurately measure incident detection through alarms on the
corridor. As a result, the evaluation of the radar’s ability to detect incidents on the corridor was
paused to allow for the adjustments to be made on the central software.
Subsequent adjustments were made to the central software in the months of October and
November 2015 to improve the accuracy of measuring vehicle counts and speeds on the I-94
corridor. These adjustments were completed in December and the evaluation of the radar
resumed in late December. To allow for further data collection on the project, the end date for
the radar’s operation was extended 3 months from Dec. 31st, 2015 out to March 31st, 2016.
While the system’s accuracy at detecting and measuring vehicle speeds and counts was
significantly increased along the I-94 corridor, it is recommended from the radar unit’s
manufacturer (Navtech) to install a different model of radar equipment that can better measure
vehicle speeds and counts on an arterial type of roadway without losing the capability of
detecting stopped vehicles for incident detection purposes.
Table 3-2 – High-Level System Installation and Adjustment Timeline
Date Activity Notes
June 2nd, 2015
Radar and cameras installed on the
pole; equipment placed into cabinet
for verifying connection to MnDOT
RTMC
Configuration of IP address for the
radar required a second trip to the
corridor; leveling of the radar with
an inclinometer took longer than
planned.
June 8th, 2015 Radar IP address configured at the
cabinet
Able to view the radar data through
central software. Central software
calibration activity performed by
Navtech staff.
15
Table 3-2 – High-Level System Installation and Adjustment Timeline
Date Activity Notes
June 17th, 2015
Radar equipment on the pole was
adjusted to allow radar to pick up
more vehicles on the east half of the
corridor. North-south tilt was
increased from 1 to 7 degrees
Adjustments were made manually
without use of inclinometer.
June 19th, 2015
Initial system configurations were
completed through central software
after visit from Navtech staff
Initial date for confirming system
operations and beginning evaluation
of radar
July 20th, 2015
Radar equipment was adjusted on
the pole to allow radar to pick up
more vehicles on the west edge of
the corridor
Adjustment with the inclinometer to
decrease the north-south tilt of the
radar that was made on June 17th
(from 7 degrees to 1 degree)
October 5th,
2015
Adjustments to rules within the
software about vehicle detection
were made to improve vehicle count
/ speed detection
Adjustments negatively impacted
the ability of the software to issue
alerts for stopped vehicles on the
corridor; evaluation of radar paused
October 23rd
through Dec. 7th
Navtech radar staff dedicated time to
improving the capability of the radar
to measure vehicle counts and
speeds at an accurate level while
maintaining the ability of the radar
to detect incidents
Multiple meetings held with
Navtech staff to understand progress
being made on the creation of new
zones on the corridor for vehicle
counts and speeds
Dec. 18th, 2015 Date set for resuming evaluation of
the radar per system requirements
Small adjustments made to zones to
improve detection of vehicles on
shoulder lanes. New set of sections
were implemented on the corridor.
January 15th,
2016
Software upgrade made. Adjustment
made to system parameters by
Navtech that negatively impacted
the ability for the software to detect
stopped vehicles measured by radar
Result of changes made to system
parameters was a sharp decrease in
the number of incidents detected by
the radar and impact to accuracy of
speed measurements.
March 2nd, 2016
Correction implemented on system
parameters to increase the ability of
the software to pick up vehicles and
incidents more accurately
Result of correction to parameters
was an increase in the number of
incidents detected and the accuracy
of incident detection and vehicle
counts
March 31st,
2016 End of operational test for project.
End date of system operations for
evaluation purposes.
16
Chapter 4 360 Degree Radar System Evaluation
This section contains the results of the evaluation of the 360 Degree Radar Information System.
In following with the Systems Engineering process, the system will be evaluated against the
goals and objectives of the project, in addition to being evaluated against the accuracy of the data
collected by the radar and processed by the software.
4.1. System Validation Process
The intent of the System Validation process is to measure the goals and objectives that were
documented within the Concept of Operations for the project. Measures of Effectiveness
(MOEs) have been proposed to quantifiably demonstrate that the system is meeting the goals and
objectives that were defined at the beginning of the project. Table 4-1 below contains a
summary of the proposed MOEs that trace back to the goals and objectives of the 360 Degree
Radar Information System previously identified in the Concept of Operations. Given that the
system was operational across multiple seasons, system operations in all types of weather
conditions can be evaluated.
Table 4-1 – Validation Plan Measures of Effectiveness
Goals and Objectives
from Concept of Operations Validation Plan MOEs
Goal
Perform traffic / incident detection and
data collection activities accurately
using the Navtech 360 Degree Radar
Information System
System positively identifies all
desired types of measures through
alerts
Objectives
1. Test the accuracy of traffic
congestion detected against camera
images monitoring traffic along the
corridor and against other sources
of data (i.e. Wavetronix data)
System positively identifies traffic
congestion through accurate
vehicle counts and alerts when
speeds are detected below pre-
determined MPH thresholds
2. Test the accuracy of incident
detection alerts recorded by the
system against other sources of
incident data (i.e. CAD system)
System positively identifies a
traffic incident through a pre-
configured alert threshold
3. Measure the efficiency of the
system’s operations in all weather
conditions to verify reliability of
system alerts generated
Determine percentage variance
between “false positive” alerts in
various weather conditions
4. Test the accuracy of all other alerts
detected against camera images
monitoring traffic along the
corridor and against other sources
of data (i.e. CAD System)
System positively identifies all
other conditions through pre-
configured alert thresholds
17
4.2. System Validation Results
A summary of the four MOEs listed in Table 4-1 is presented in the following sub-sections.
These sections were included within the Validation Plan document developed for the project and
have been modified to reflect actual conditions encountered in the project.
4.2.1. MOE #1 – Speed Threshold Alerts and Vehicle Counts
This MOE pertains to the accuracy of alerts generated when vehicle speeds are detected
below a pre-determined speed threshold. The speed threshold has been set at 15 miles
per hour (MPH). An alert is generated in the system and logged when a vehicle’s speed
is detected below this threshold.
The speed threshold data generated by the 360 degree radar system has been validated
against lane speeds measured from the Wavetronix detector installed within the project
area. Detector data has also been gathered from the MnDOT DataExtract Tool at 1-
minute intervals that indicate the measured travel speeds. DataExtract provides lane-
specific travel speeds have been used to validate the speed threshold alerts generated by
the 360 Degree Radar System.
DataExtract has also been used in a similar manner to validate the accuracy of vehicle
counts by the 360 Degree Radar Information System. DataExtract provides lane-specific
vehicle counts that are also used to validate the accuracy of vehicle counts down to a 1-
minute interval. DataExtract exports the vehicle count data into an MS Excel file for ease
of comparing count data in MS Excel.
During the course of the demonstration, there were a large amount of speed threshold
alerts generated from slow moving travel in the rush hour periods. The alerts that could
be presented for vehicle speeds less than 15 MPH was not found to be as valuable as
stopped vehicles in the corridor. In the second period of the demonstration, a sample of
speed threshold alerts was extracted from the central software for comparison with
Wavetronix data at the same time period.
The MOE was validated as the system positively identified events when speeds are
detected below the pre-determined 15 MPH threshold for all lanes of traffic in the
westbound and eastbound directions, as compared against the MnDOT DataExtract that
was also gathered.
Regarding vehicle counts, this MOE was validated as part of the evaluation of the vehicle
counts measured against Wavetronix data which is presented in Section 4.3.
4.2.2. MOE #2 – Traffic Incident Alerts
This MOE pertains to the accuracy of alerts generated when stopped vehicles are detected
within the project area, which may include traffic incidents involving one or multiple
vehicles that require emergency response.
The process of validating traffic incident alerts involves verifying those incidents against
incident logs made by RTMC operations staff. Figure 4-1 below presents a screenshot of
18
an incident from Dec. 17th, 2015 in the westbound direction of the corridor limits. The
incident notification time was logged as 12:57 pm. These incidents are reviewed with
MnDOT cameras to confirm that the events occurred within the project limits of the
corridor. A screenshot of the camera confirmation is presented in Figure 4-2, with the
timestamp and incident also highlighted.
Figure 4-1 – MnDOT Computer Aided Dispatch Summary of Dec. 17th Event
Figure 4-2 – MnDOT Camera Record of Dec. 17th Event for Confirmation with
360 Degree Radar System
Location of stopped vehicles
Timestamp of incident in video playback
Timestamp of incident notification in MnDOT CAD
MnDOT CAD EID 7750376
19
The 360 Degree Radar system recorded the event as shown in Figure 4-3. The
timestamp that aligns best with the MnDOT incident notification is circled and
highlighted in blue. At this point in time, the vehicles shown in Figure 4-2 first
proceeded in Section 0 that represents the emergency shoulder area installed in this part
of the corridor. However, the 360 radar first recorded the stopped vehicles in Section 12
at 12:50 pm, approximately 7 minutes before received a notification from the motorists
via cell phone. This event is also circled in Figure 4-3.
Figure 4-3 – 360 Degree Radar System Record of Dec. 17th Event
The notification of the incident at 12:50pm is also recorded by the CCTV cameras in the
NVR. This is shown in Figure 4-4. The location of the incident in recorded by the radar
in Section 12 is circled in the still image. Note that the incident was recorded at 12:50:32
in the timestamp of the camera. The 360 Degree Radar central software was configured
to record an event after a vehicle was stopped for a period of 10 seconds in the corridor.
The timestamp of the alert in Figure 4-3 is recorded about 10 seconds later at 12:50:41.
The two vehicles involved in the incident then pulled ahead into the emergency shoulder
area at about 12:58 with the assistance of a MnDOT FIRST vehicle that arrived on the
scene. As the vehicles stopped in that area, an additional alert was created in the
software, which can also be seen in Figure 4-3.
Timestamp of stopped vehicles match with MnDOT incident notification
Timestamp of stopped vehicle alert in 360 radar system – 7 minutes prior to
MnDOT notification in CAD system
20
Incident recorded on camera by NVR
Camera timestamp NVR timestamp
Figure 4-4 – Interface of NVR with Video Record Dec. 17th Event
This MOE has been considered as validated through the comparison of 360 Degree Radar
system events, as compared against CCTV camera footage recorded on the NVR unit and
against MnDOT CAD events logged throughout the project. A summary of the events
compared against the MnDOT CAD system is presented on the following pages.
Table 4-2 presents a summary of events from the August to September period of incident
data collection. Table 4-3 presents a summary of events from the February to March
period of incident data collection.
It should be noted that various updates to the central software occurred during the periods
of October and November that limited the ability of the system to accurately detect
incidents in the project area. In addition, it was learned through the course of the project
that heavy traffic congestion in the project area caused the software to ignore multiple
alarms of stopped vehicles at the same time. This had the impact of causing the software
to ignore stopped vehicle events reported by the radar unit that were actual incidents
logged by the RTMC operators. Navtech Radar has learned of this issue through this
project and is actively developing a software upgrade to address this issue in future
deployments. These events of congestion (i.e. stopped vehicles) masking the traffic
incidents are noted in Tables 4-2 and 4-3 below under the Description column.
21
Table 4-4 presents a summary of 11 events from the August to September period that
were un-detected by RTMC operators, but were recorded the 360 Degree Radar
Information System. This finding for the two month period of time indicates that other
events may have gone un-detected by RTMC operators throughout the course of the
project because those events were not called in to 911 operators as quickly as they could
have been by those involved in the incidents.
22
Table 4-2 – Summary of Events Logged by 360 Degree Radar and MnDOT CAD System (August – September 2015)
Date Radar Alarm
Radar Time
Section CAD Event CAD Time Radar/CAD Time Diff.
CAD EID CAD
Response Response
Time Description
8/9/2015 Stopped vehicle
15:03 2 Crash 3:03 PM 0 min 7479318
8/11/2015 Stopped vehicle
13:15 10 Crash 1:22 PM +7 min 7482760 FIRST 1:34 PM Tow dispatched at 1:34 pm. Trooper dispatched at 1:44 pm.
8/12/2015 None Stall 4:59 PM 7485217 none -- General congestion that masked an incident by radar.
8/19/2015 Stopped vehicle
13:00 2 Crash 1:03 PM +3 min 7499194 FIRST 1:09 PM
Crash in section 2, vehicles stopped in section 3. Tow vehicle dispatched at 1:23pm as well.
8/25/2015 Stopped vehicle
14:27 10 Stall 2:44 PM +17 min 7512696 FIRST 2:44 PM Trucked pulled over and stopped.
9/4/2015 Stopped vehicle
17:56 7 Stall 5:59 PM +3 min 7537516 none Bus pulled over on shoulder.
9/4/2015 Stopped vehicle
18:50 10 Crash 6:53 PM +3 min 7537624 FIRST 6:56 PM Two cars involved in crash.
9/9/2015 None Stall 9:40 AM 7547724 FIRST 9:40 AM General congestion that masked an incident by radar.
9/14/2015 Stopped vehicle
19:19 7 Stall 7:19 PM 0 min 7559448 FIRST 7:19 PM
9/15/2015 None Stall 4:40 PM 7561197 FIRST 5:27 PM General congestion that masked an incident by radar.
9/19/2015 Stopped vehicle
12:19 10 Crash 12:29 PM +10 min 7568883 Trooper 12:53 PM Two cars involved in crash.
Average Difference of Notification Time (Based on 8 events in bold) + 5.4 min
23
Table 4-3 – Summary of Events Logged by 360 Degree Radar and MnDOT CAD System (February - March 2016)
Date Radar Alarm
Radar Time
Section CAD
Event CAD Time
Radar/CAD Time Diff.
CAD EID CAD
Response Response
Time Description
12/17/2015 Stopped vehicle
10:30 2 Stall 10:38 AM +8 Min. 7750107 none none Tractor trailer stalled.
12/17/2015 Stopped vehicle
12:50 10 Crash 12:57 PM +7 Min. 7750376 FIRST 12:57 PM
Actual crash recorded. Drivers got out and stood on highway. Image shown in Figure 4-4.
2/29/2016 Stopped vehicle
7:22 0 Crash 8:20 AM +58 Min. 7892574 FIRST +
Tow Truck
8:27 AM + 8:40AM
3/1/2016 None n/a 3 Crash 7:39 AM n/a 7894412
Trooper + FIRST +
Tow Truck
7:50 AM + 7:56 AM + 8:04 AM
Eastbound in constant queue.
3/4/2016 Stopped vehicle
16:11 3 Stall 4:01 PM n/a 7901374 FIRST 4:10 PM
General congestion that masked an incident by radar. Detected MnDOT vehicle @ 4:11pm.
3/22/2016 Stopped vehicle
17:29 0 Crash 5:12 PM n/a 7938652 FIRST + Trooper
5:23 PM + 5:26 PM
Congestion Queue activated during this time
3/25/2016 None n/a 2 Crash 4:25 PM n/a 7945331 FIRST 4:25 PM General congestion that masked an incident by radar.
Average Difference of Notification Time (Based on 3 events in bold) + 24.3 min
Average Difference of Notification Time for all Incidents in Project (Based on 11 total events in bold in Table 4-2 and Table 4-3)
+ 10.54 minutes
24
Table 4-4 – Summary of Events Logged by 360 Degree Radar and MnDOT CAD System
Radar Events Not On CAD with Minimum of 4 Minute Alarm (Total of 12 events listed)
Date Radar Alarm Radar Time Section
CAD Event Description
8/9/2015 Stopped vehicle 10:23 10 None Vehicle in pulloff for 8 minutes
8/10/2015 Stopped vehicle 15:58 2 None 2 vehicles pulled onto shoulder, got out of vehicles, walked around
8/11/2015 Stopped vehicle 4:59 7 None Vehicle stopped on shoulder for a minimum of 5 minutes.
8/12/2015 Stopped vehicle 10:30 7 None Pickup truck stopped on shoulder for a minimum of 5 minutes.
8/12/2015 Stopped vehicle 14:05 10 None Truck pulled over and stopped for a minimum of 7 minutes.
8/15/2015 Stopped vehicle 17:45 10 None Trucked pulled over and stopped for a minimum of 22 minutes, possibly 1 hour.
8/19/2015 Stopped vehicle 1:30 10 None 2 cars pulled over and stopped for a minimum of 5 minutes.
8/20/2015 Stopped vehicle 11:25 10 None 2 cars pulled over and stopped for a minimum of 4 minutes.
8/20/2015 Stopped vehicle 20:24 10 None Truck pulled over and stopped for a minimum of 5 minutes.
8/25/2015 Stopped vehicle 14:44 10 None Truck pulled over and stopped for a minimum of 4 minutes.
9/1/2015 Stopped vehicle 9:56 10 None Pickup truck stopped on shoulder for a minimum of 8 minutes.
12/18/2015 Stopped vehicle 9:58 0 None Vehicle pulled off for over 10 minutes, person got out and walked around.
Notes: 1) There were 10 to 15 more alarms ranging from 1 to 2 minutes where the vehicles were most likely stopped for a longer period of time.
2) A stopped vehicle will only alarm for 1 to 2 minutes then "blend" it into the landscape even if it is stopped for +2 minutes. This helps to eliminate
"blobs" on the radar detection. We are working on trying to track stopped vehicles for a longer period of time. Alarms for large trucks last longer as they
take longer to "blend" into the landscape. This is one reason we see more alarms for trucks.
25
4.2.3. MOE #3 – Measure of “False Positive” Alerts
This MOE pertains to the number of “false positive” alerts that may be generated by
the system for all types of pre-configured alerts.
A “false positive” alert is defined as an alert that cannot be verified against a camera
recording on the NVR unit of an event for which the alert was created, or against data
gathered from the MnDOT DataExtract for the purposes of verifying the alert was
accurate.
Based on prior project deployments, the manufacturer of the radar has noted that false
positives are likely to occur in the initial stages of the radar’s operation. As these
false positive alerts are reviewed, the occurrence can be greatly reduced by modifying
the rules within the software that are set to define when alerts should be generated.
Table 4-5 illustrates the false alarms observed for the period of June 20th to June 26th
after initial system configuration. A total of 5 false alarms were noted (out of a total
of 41 events for the period), with descriptions and possible causes listed. The average
false alarm rate over these 7 days was recorded 0.71, which would be less than one
per day.
Table 4-5 – Summary of False Alarms Logged June 20th to June 26th
Date Time Alarm Type
East / West
Section Alarm
(Verified / False)
Description Possible Cause
6/22/2015 16:53:38 Stopped vehicle
East-bound
6 False Shows tracks over all 3 lanes. 35W ramp moving slow with truck passing.
6/23/2015 7:39:07 Stopped vehicle
East-bound
7 False
Shows tracks on shoulder close to off-ramp. Other tracks shown around same time. 35W on-ramp stopped.
Per video truck stopped on 35W on-ramp, possibly causing false alarms.
6/23/2015 8:09:23 Stopped vehicle
East-bound
6 False
Shows tracks on shoulder close to off-ramp. Other tracks shown around same time. 35W on-ramp stopped.
Per video 2 large semi's pass slowly during alarm on 35W on-ramp possibly causing false alarm.
6/26/2015 12:56:09 Reverse vehicle
East-bound
7 False No vehicle in reverse. Westbound traffic very slow.
6/26/2015 15:38:11 Reverse vehicle
East-bound
6 False No vehicle in reverse.
Westbound traffic very slow, truck passing slowly westbound.
26
Table 4-6 illustrates the false alarms observed for the period of March 2nd to March
9th, 2016 after central software updates were made later in the project. A total of 2
false alarms were noted (out of a total of 66 events for the period), with descriptions
and possible causes listed. The average false alarm rate over these 7 days was
recorded 0.28, which would be less than one false alarm per day. This decrease in the
false alarm rate observed from the beginning of the project to the end of the project
helps to demonstrate how the accuracy of incident detection by the 360 Degree Radar
Information System improves over time as the system is updated to better understand
how to ignore previous false alarms encountered in a project.
Table 4-6 – Summary of False Alarms Logged March 2nd to March 9th
Date Time Alarm Type
East / West
Section Alarm
(Verified / False)
Description Possible Cause
3/3/2016 1:25:16 Stopped vehicle
East-bound
3 False No stopped vehicle present
3/3/2016 1:25:16 Stopped vehicle
East-bound
3 False No stopped vehicle present
4.2.4. MOE #4 – All Other Condition Alerts
This MOE pertains to the accuracy of all other types alerts generated when pre-
determined conditions are detected within the project area. These conditions will
include the following:
1. Individual person(s) detected within the project area
2. Vehicles slowly moving in reverse on the corridor
For item #1, an alert is generated in the system and logged when a person is detected
within the project area. These alerts can be categorized in a separate Microsoft Excel
file for review in analysis of the system alerts. Video is recorded of the person(s) on
the NVR unit within the system server for verification purposes. The timestamp of
the video recording is then correlated with the timestamp of the alert recorded on the
system server.
The alerts are validated when the video recording shows the person(s) within the field
project area. This portion of the MOE has been validated based on events logged by
the radar compared against CCTV camera footage recorded on the NVR unit.
For item #2, an alert is generated in the system and logged when vehicles are detected
to be slowly moving in the reverse direction on any portion of the corridor. This
portion of the MOE has been validated based on the reverse movement of vehicles
detected on the corridor, as compared against CCTV camera footage recorded on the
NVR unit.
27
Table 4-7 illustrates the number of alarms that were “verified” for the period of June
20th to June 26th after initial system configuration. A total of 6 verified “person”
alarms and 2 “reverse vehicle” alarms were confirmed against CCTV recorded video,
with descriptions listed. Some of the verified alarms relate to the same event that
occurred within the field.
Table 4-7 – Summary of Person and Reverse Vehicle Alarms Verified June 20th to June 26th
Date Time Alarm Type East/West Section Description
6/20/2015 15:06:54 Person Westbound 10 4 people get out of 3 vehicles on pulloff.
6/20/2015 22:02:48 Reverse vehicle
Eastbound 6 Vehicle reversed on shoulder 300ft to get on 35W on-ramp.
6/21/2015 4:21:29 Reverse vehicle
Eastbound 6 MnDOT FIRST vehicle backing up on fast lane shoulder.
6/22/2015 11:55:30 Person Westbound 2 Driver and other vehicle came back to get vehicle.
6/22/2015 11:59:39 Person Westbound 2 Same driver causing another alarm.
6/22/2015 19:20:36 Person Westbound 2 Tow truck driver.
6/25/2015 0:53:59 Person Westbound 2 Road crew got out of truck.
6/26/2015 7:54:01 Person Westbound 10 Person walked around truck.
4.3. System Evaluation Process
For the more detailed System Evaluation, a subset of data was gathered at various 15-minute
time periods over the course of multiple dates. A summary of the time periods is presented
in Table 4-8.
Table 4-8 – Dates of Vehicle Speed and Count Comparisons with Wavetronix Data
Dates of Data Collection Time Periods Notes
Nov. 19th – 20th, 2015
2:30 am to 2:45 am
10:00 am to 10:15am
2:00 pm to 2:15 pm
5:00 pm to 5:15 pm
Initial comparison of vehicle
speeds and counts with the
Wavetronix Data. Manual
counts also performed.
Jan. 17th, 2016
3:30 am to 3:45 am
9:30 am to 9:45 am
1:30 pm to 1:45 pm
5:30 pm to 5:45 pm
High temperature of minus 8
degrees Fahrenheit.
Feb. 2nd, 2016
3:15 am to 3:30 am
11:15 am to 11:30 am
1:15 pm to 1:30 pm
5:15 pm to 5:30 pm
Heavy snowfall (10 inches)
from 12 noon to 12 midnight.
28
Radar data and CCTV camera video were recorded on the NVR unit for playback and
evaluation purposes against data gathered from the MnDOT DataExtract tool over an
identical period of time.
Vehicle speeds and vehicle counts measured by the 360 Degree Radar System were
compared against vehicle speeds estimated by the Wavetronix unit installed at the project
area measuring each lane of traffic. Westbound lanes are indicated as #356W1, #357W2,
and #358W3, while the eastbound lanes are indicated as #497E1, #498E2, and #499E3. For
the Nov. 20th date of data collection, visual counts of traffic using CCTV camera recordings
of traffic were performed to have an independent point of comparison with both the 360
Degree Radar and the Wavetronix data sets.
Data on each lane of travel was gathered from the DataExtract tool. Vehicle speeds
estimated by the Wavetronix unit were gathered at 1-minute intervals during the time periods
indicated in Table 4-8. Vehicle speed data for each lane of traffic, as well as an Aggregate
measure of all lanes, was gathered from the DataExtract tool and exported to a Microsoft
Excel file for ease of analysis.
The Excel files were imported into a separate Microsoft Excel file generated through use of
the 360 Degree Radar central software. Charts and graphs were created in Microsoft Excel to
compare radar data with the MnDOT data to determine how close the two sets of data were
to each other.
4.4. System Evaluation Results
The System Evaluation results are presented in the following Figures in this section.
Multiple dates are presented in the following figures that are summarized below. For data
collected on Thursday, Nov. 19th and Friday, Nov. 20th, 2015, the following figures have
been created on the following pages:
Figure 4-5 – Navtech and Wavetronix westbound (WB) and eastbound (EB) speed
comparisons for all lanes of traffic combined based on data collected on Nov. 20th,
2015 at three different time periods
Figure 4-6 – Navtech and Wavetronix westbound (WB) and eastbound (EB) speed
comparisons for individual lanes of traffic based on data collected on Nov. 20th, 2015
at different time periods
Figures 4-7 – Navtech, Wavetronix, and Manual vehicle count comparisons for
westbound (WB) and eastbound (EB) directions based on data collected on Nov. 19th,
2015 at four different time periods
Figure 4-8 – Navtech and Wavetronix westbound (WB) and eastbound (EB) vehicle
count comparisons for individual lanes of traffic based on data collected on Nov. 19th,
2015 at different time periods
For data collected on Sunday, Jan. 17th, 2016, the figures listed have been created on the
following pages. It should be noted that this date was selected for the very low temperatures
observed. The high temperature recorded on this date was minus 2 degrees Fahrenheit, with
the low temperature at minus 13 degrees Fahrenheit. Winds between 10 and 20 miles per
29
hour were recorded, which created wind chill temperatures lower than the noted low
temperature of minus 13 degrees.
Figure 4-9 – Navtech and Wavetronix westbound (WB) and eastbound (EB) speed
comparisons for all lanes of traffic combined based on data collected on Jan. 17th,
2016 at different time periods
Figure 4-10 – Navtech and Wavetronix westbound (WB) and eastbound (EB) speed
comparisons for individual lanes of traffic based on data collected on Jan. 17th, 2016
at different time periods
Figures 4-11 – Navtech and Wavetronix vehicle count comparisons for westbound
(WB) and eastbound (EB) directions based on data collected on Jan. 17th, 2016 at
different time periods
Figure 4-12 – Navtech and Wavetronix westbound (WB) and eastbound (EB) vehicle
count comparisons for individual lanes of traffic based on data collected on Jan. 17th,
2016 at different time periods
For data collected on Tuesday, Feb. 2nd, 2016, the figures listed have been created on the
following pages. It should be noted that this date was selected for the high amount of
snowfall throughout the day and the low temperatures observed. Approximately 10 inches of
snow fell on this date in the Minneapolis area between 12 noon and midnight.
Figure 4-13 – Navtech and Wavetronix westbound (WB) and eastbound (EB) speed
comparisons for all lanes of traffic combined based on data collected on Feb. 2nd,
2016 at different time periods
Figure 4-14 – Navtech and Wavetronix westbound (WB) and eastbound (EB) speed
comparisons for individual lanes of traffic based on data collected on Feb. 2nd, 2016
at different time periods
Figures 4-15 – Navtech and Wavetronix vehicle count comparisons for westbound
(WB) and eastbound (EB) directions based on data collected on Feb. 2nd, 2016 at
different time periods
Figure 4-16 – Navtech and Wavetronix westbound (WB) and eastbound (EB) vehicle
count comparisons for individual lanes of traffic based on data collected on Feb. 2nd,
2015 at different time periods
For all three time periods observed, vehicle counts remained relatively close to the
Wavetronix unit on the pole at the location of the radar unit regardless of weather conditions
observed on these dates. The vehicle speed measurements were also relatively close for the
November time period, but were consistently lower than what the Wavetronix device had
reported in the January and February time periods. This is primarily due to changes that
were made to system parameters on Jan. 15th that negatively impacted the accuracy of speed
measurements, as well as a decrease in the number of incidents recorded after this date.
30
Figure 4-5 – Navtech and Wavetronix WB and EB Speed Comparisons on Nov. 20th
31
Figure 4-5 (continued) – Navtech and Wavetronix WB and EB Speed Comparisons on Nov. 20th
32
Figure 4-5 (continued) – Navtech and Wavetronix WB and EB Speed Comparisons on Nov. 20th
33
Figure 4-6 – Navtech and Wavetronix WB and EB Speed Comparisons on Nov. 20th
By Lane for all Time Periods
34
Figure 4-7 – Navtech, Wavetronix, and Manual WB and EB Count Comparisons on Nov. 19th
35
Figure 4-7 (continued) – Navtech, Wavetronix, and Manual WB and EB Count Comparisons on Nov. 19th
36
Figure 4-7 (continued) – Navtech, Wavetronix, and Manual WB and EB Count Comparisons on Nov. 19th
37
Figure 4-7 (continued) – Navtech, Wavetronix, and Manual WB and EB Count Comparisons on Nov. 19th
38
Figure 4-8 – Navtech and Wavetronix WB and EB Count Comparisons on Nov. 19th By
Lane for all Time Periods
39
Figure 4-9 – Navtech and Wavetronix WB and EB Speed Comparisons on Jan. 17th, 2016
0
10
20
30
40
50
60
70
80
9:30 AM 9:33 AM 9:36 AM 9:38 AM 9:41 AM 9:44 AM 9:47 AM
Spee
d (
MP
H)
Time
WB Speed Comparison (All Lanes) -- Jan. 17th, 2016 - 9:30 am to 9:45 am
Navtech
Wavetronix
0
10
20
30
40
50
60
70
80
9:30 AM 9:33 AM 9:36 AM 9:38 AM 9:41 AM 9:44 AM 9:47 AM
Spee
d (
MP
H)
Time
EB Speed Comparison (All Lanes) -- Jan. 17th, 2016 - 9:30 am to 9:45 am
Navtech
Wavetronix
40
Figure 4-9 (continued) – Navtech and Wavetronix WB and EB Speed Comparisons on Jan. 17th, 2016
0
10
20
30
40
50
60
70
1:29 PM 1:32 PM 1:35 PM 1:37 PM 1:40 PM 1:43 PM 1:46 PM
Spee
d (
MP
H)
Time
WB Speed Comparison (All Lanes) -- Jan. 17th, 2016 - 1:30 pm to 1:45 pm
Navtech
Wavetronix
0
10
20
30
40
50
60
70
1:29 PM 1:32 PM 1:35 PM 1:37 PM 1:40 PM 1:43 PM 1:46 PM
Spee
d (
MP
H)
Time
EB Speed Comparison (All Lanes) -- Jan. 17th, 2016 - 1:30 pm to 1:45 pm
Navtech
Wavetronix
41
Figure 4-9 (continued) – Navtech and Wavetronix WB and EB Speed Comparisons on Jan. 17th, 2016
0
10
20
30
40
50
60
70
5:28 PM 5:31 PM 5:34 PM 5:36 PM 5:39 PM 5:42 PM 5:45 PM 5:48 PM
Spee
d (
MP
H)
Time
WB Speed Comparison (All Lanes) -- Jan. 17th, 2016 - 5:30 pm to 5:45 pm
Navtech
Wavetronix
0
10
20
30
40
50
60
70
5:28 PM 5:31 PM 5:34 PM 5:36 PM 5:39 PM 5:42 PM 5:45 PM 5:48 PM
Spee
d (
MP
H)
Time
EB Speed Comparison (All Lanes) -- Jan. 17th, 2016 - 5:30 pm to 5:45 pm
Navtech
Wavetronix
42
Figure 4-10 – Navtech and Wavetronix WB and EB Speed Comparisons on Jan. 17th
By Lane for all Time Periods
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
0315-0330 0915-0930 1315-1330 1715-1730
Per
cen
tage
Dif
fere
nce
Time
Navtech WB Average Speed Comparison Per Lane with Wavetronix Data Jan. 17th, 2016
Slow Lane Middle Lane Fast Lane
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
0315-0330 0915-0930 1315-1330 1715-1730
Per
cen
tage
Dif
fere
nce
Time
Navtech EB Average Speed Comparison Per Lane with Wavetronix Data Jan. 17th, 2016
Slow Lane Middle Lane Fast Lane
43
Figure 4-11 – Navtech and Wavetronix WB and EB Count Comparisons on Jan. 17th
0
2
4
6
8
10
12
14
16
3:30 AM 3:33 AM 3:36 AM 3:38 AM 3:41 AM 3:44 AM 3:47 AM
Co
un
t
Time
EB Count Comparison (All Lanes) -- Jan. 17th, 2016 - 3:30 am to 3:45am
Navtech
Wavetronix
0
10
20
30
40
50
60
9:30 AM 9:33 AM 9:36 AM 9:38 AM 9:41 AM 9:44 AM 9:47 AM
Co
un
t
Time
EB Count Comparison (All Lanes) -- Jan. 17th, 2016 - 9:30am to 9:45 am
Navtech
Wavetronix
44
Figure 4-11 (continued) – Navtech and Wavetronix WB and EB Count Comparisons on Jan. 17th
0
20
40
60
80
100
120
1:29 PM 1:32 PM 1:35 PM 1:37 PM 1:40 PM 1:43 PM 1:46 PM
Co
un
t
Time
EB Count Comparison (All Lanes) -- Jan. 17th, 2016 - 1:30 pm to 1:45 pm
Navtech
Wavetronix
0
10
20
30
40
50
60
70
80
5:28 PM 5:31 PM 5:34 PM 5:36 PM 5:39 PM 5:42 PM 5:45 PM 5:48 PM
Co
un
t
Time
EB Count Comparison (All Lanes) -- Jan. 17th, 2016 - 5:30 pm to 5:45 pm
Navtech
Wavetronix
45
Figure 4-11 (continued) – Navtech and Wavetronix WB and EB Count Comparisons on Jan. 17th
0
2
4
6
8
10
12
14
3:30 AM 3:33 AM 3:36 AM 3:38 AM 3:41 AM 3:44 AM 3:47 AM
Co
un
t
Time
WB Count Comparison (All Lanes) -- Jan. 17th, 2016 - 3:30 am to 3:45 am
Navtech
Wavetronix
0
5
10
15
20
25
30
35
40
45
50
9:30 AM 9:33 AM 9:36 AM 9:38 AM 9:41 AM 9:44 AM 9:47 AM
Co
un
t
Time
WB Count Comparison (All Lanes) -- Jan. 17th, 2016 - 9:30 am to 9:45am
Navtech
Wavetronix
46
Figure 4-11 (continued) – Navtech and Wavetronix WB and EB Count Comparisons on Jan. 17th
0
10
20
30
40
50
60
70
80
1:29 PM 1:32 PM 1:35 PM 1:37 PM 1:40 PM 1:43 PM 1:46 PM
Co
un
t
Time
WB Count Comparison (All Lanes) -- Jan. 17th, 2016 - 1:30 pm to 1:45 pm
Navtech
Wavetronix
0
10
20
30
40
50
60
70
80
90
100
5:28 PM 5:31 PM 5:34 PM 5:36 PM 5:39 PM 5:42 PM 5:45 PM 5:48 PM
Co
un
t
Time
WB Count Comparison (All Lanes) -- Jan. 17th, 2016 - 5:30 pm to 5:45 pm
Navtech
Wavetronix
47
Figure 4-12 – Navtech and Wavetronix WB and EB Count Comparisons on Jan. 17th
By Lane for all Time Periods
-40
-20
0
20
40
60
80
100
120
0315-0330 0915-0930 1315-1330 1715-1730
Per
cen
tage
Dif
fere
nce
Time
Navtech WB Count Comparison Per Lane with Wavetronix Data Jan. 17th, 2016
Slow Lane Middle Lane Fast Lane
-150
-100
-50
0
50
100
150
0315-0330 0915-0930 1315-1330 1715-1730
Per
cen
tage
Dif
fere
nce
Time
Navtech EB Count Comparison Per Lane with Wavetronix Data Jan. 17th, 2016
Slow Lane Middle Lane Fast Lane
48
Figure 4-13 – Navtech and Wavetronix WB and EB Speed Comparisons on Feb. 2nd, 2016
0
10
20
30
40
50
60
70
11:13 AM 11:16 AM 11:19 AM 11:22 AM 11:25 AM 11:28 AM 11:31 AM
Spee
d (
MP
H)
Time
WB Speed Comparison (All Lanes) -- Feb. 2nd, 2016 - 11:15 am to 11:30 am
Navtech
Wavetronix
0
10
20
30
40
50
60
70
11:13 AM 11:16 AM 11:19 AM 11:22 AM 11:25 AM 11:28 AM 11:31 AM
Spee
d (
MP
H)
Time
EB Speed Comparison (All Lanes) -- Feb. 2nd, 2016 - 11:15 am to 11:30 am
Navtech
Wavetronix
49
Figure 4-13 (continued) – Navtech and Wavetronix WB and EB Speed Comparisons on Feb. 2nd, 2016
0
2
4
6
8
10
12
14
16
18
1:14 PM 1:17 PM 1:20 PM 1:23 PM 1:26 PM 1:29 PM 1:32 PM
Spee
d (
MP
H)
Time
WB Speed Comparison (All Lanes) -- Feb. 2nd, 2016 - 1:15 pm to 1:30 pm
Navtech
Wavetronix
0
5
10
15
20
25
30
35
40
45
1:14 PM 1:17 PM 1:20 PM 1:23 PM 1:26 PM 1:29 PM 1:32 PM
Spee
d (
PM
H)
Time
EB Speed Comparison (All Lanes) -- Feb. 2nd, 2016 - 1:15 pm to 1:30 pm
Navtech
Wavetronix
50
Figure 4-13 (continued) – Navtech and Wavetronix WB and EB Speed Comparisons on Feb. 2nd, 2016
0
5
10
15
20
25
5:13 PM 5:16 PM 5:19 PM 5:22 PM 5:25 PM 5:28 PM 5:31 PM
Spee
d (
MP
H)
Time
WB Speed Comparison (All Lanes) -- Feb. 2nd, 2016 - 5:15 pm to 5:30 pm
Navtech
Wavetronix
0
5
10
15
20
25
30
35
40
45
5:13 PM 5:16 PM 5:19 PM 5:22 PM 5:25 PM 5:28 PM 5:31 PM
Spee
d (
MP
H)
Time
EB Speed Comparison (All Lanes) -- Feb. 2nd, 2016 - 5:15 pm to 5:30 pm
Navtech
Wavetronix
51
Figure 4-14 – Navtech and Wavetronix WB and EB Speed Comparisons on Feb. 2nd
By Lane for all Time Periods
-40
-35
-30
-25
-20
-15
-10
-5
0
0230-0245 1000-1015 1400-1415 1700-1715
Per
cen
tage
Dif
fere
nce
Time
Navtech WB Average Speed Comparison Per Lane with Wavetronix Data Feb. 2nd, 2016
Slow Lane Middle Lane Fast Lane
-80
-70
-60
-50
-40
-30
-20
-10
0
0230-0245 1000-1015 1400-1415 1700-1715
Per
cen
tage
Dif
fere
nce
Time
Navtech EB Average Speed Comparison Per Lane with Wavetronix Data Feb. 2nd, 2016
Slow Lane Middle Lane Fast Lane
52
Figure 4-15 – Navtech and Wavetronix WB and EB Count Comparisons on Feb. 2nd
0
2
4
6
8
10
12
14
16
3:12 AM 3:15 AM 3:18 AM 3:21 AM 3:24 AM 3:27 AM 3:30 AM 3:33 AM
Co
un
t
Time
EB Count Comparison (All Lanes) -- Feb. 2nd, 2016 - 3:15 am to 3:30 am
Navtech
Wavetronix
0
10
20
30
40
50
60
70
80
90
11:13 AM 11:16 AM 11:19 AM 11:22 AM 11:25 AM 11:28 AM 11:31 AM
Co
un
t
Time
EB Count Comparison (All Lanes) -- Feb. 2nd, 2016 - 11:15 am to 11:30 am
Navtech
Wavetronix
53
Figure 4-15 (Continued) – Navtech and Wavetronix WB and EB Count Comparisons on Feb. 2nd
0
10
20
30
40
50
60
70
1:14 PM 1:17 PM 1:20 PM 1:23 PM 1:26 PM 1:29 PM 1:32 PM
Co
un
t
Time
EB Count Comparison (All Lanes) -- Feb. 2nd, 2016 - 1:15pm to 1:30 pm
Navtech
Wavetronix
0
10
20
30
40
50
60
70
5:13 PM 5:16 PM 5:19 PM 5:22 PM 5:25 PM 5:28 PM 5:31 PM
Co
un
t
Time
EB Count Comparison (All Lanes) -- Feb. 2nd, 2016 - 5:15 pm to 5:30 pm
Navtech
Wavetronix
54
Figure 4-15 (Continued) – Navtech and Wavetronix WB and EB Count Comparisons on Feb. 2nd
0
1
2
3
4
5
6
7
8
3:12 AM 3:15 AM 3:18 AM 3:21 AM 3:24 AM 3:27 AM 3:30 AM 3:33 AM
Co
un
t
Time
WB Count Comparison (All Lanes) -- Feb. 2nd, 2016 - 3:15 am to 3:30 am
Navtech
Wavetronix
0
10
20
30
40
50
60
70
80
90
100
11:13 AM 11:16 AM 11:19 AM 11:22 AM 11:25 AM 11:28 AM 11:31 AM
Co
un
t
Time
WB Count Comparison (All Lanes) -- Feb. 2nd, 2016 - 11:15 am to 11:30 am
Navtech
Wavetronix
55
Figure 4-15 (Continued) – Navtech and Wavetronix WB and EB Count Comparisons on Feb. 2nd
0
10
20
30
40
50
60
70
80
1:14 PM 1:17 PM 1:20 PM 1:23 PM 1:26 PM 1:29 PM 1:32 PM
Co
un
t
Time
WB Count Comparison (All Lanes) -- Feb. 2nd, 2016 - 1:15 pm to 1:30 pm
Navtech
Wavetronix
0
10
20
30
40
50
60
70
5:13 PM 5:16 PM 5:19 PM 5:22 PM 5:25 PM 5:28 PM 5:31 PM
Co
un
t
Time
WB Count Comparison (All Lanes) -- Feb. 2nd, 2016 - 5:15 pm to 5:30 pm
Navtech
Wavetronix
56
Figure 4-16 – Navtech and Wavetronix WB and EB Count Comparisons on Feb. 2nd
By Lane for all Time Periods
-20
-10
0
10
20
30
40
50
60
70
0230-0245 1000-1015 1400-1415 1700-1715
Per
cen
tage
Dif
fere
nce
Time
Navtech WB Count Comparison Per Lane with Wavetronix Data Feb. 2nd, 2016
Slow Lane Middle Lane Fast Lane
-40
-20
0
20
40
60
80
0230-0245 1000-1015 1400-1415 1700-1715Per
cen
tage
Dif
fere
nce
Time
Navtech EB Count Comparison Per Lane with Wavetronix Data Feb. 2nd, 2016
Slow Lane Middle Lane Fast Lane
57
Chapter 5 Conclusions and Recommendations
This section contains conclusions on the overall 360 Radar Information System and
recommendations for MnDOT to consider in deploying further systems at other locations
along corridors and at intersections as well.
5.1. Conclusions
A summary of the goals and objectives as presented in the Concept of Operations in addition
to general conclusions on each of these objectives are provided in Table 5-1. A summary of
key findings from the evaluation are presented in Table 5-2. A summary of the operational
needs addressed by the system (as noted in Chapter 1 of this report) are presented in Table 5-
3.
Table 5-1 – Summary of Conclusions on 360 Degree Radar Information System
Goals and Objectives
from Concept of Operations Conclusions
Goal
Perform traffic / incident detection
and data collection activities
accurately using the Navtech 360
Degree Radar Information System
Low number of incident false alarms
observed at beginning of project
decreased to a smaller number near
the end of the project
Objectives
1. Test the accuracy of traffic
congestion detected against
camera images monitoring traffic
along the corridor and against
other sources of data (i.e.
Wavetronix data)
Vehicle counts and speeds from radar
unit were found to be consistent with
Wavetronix counts and speeds, as
well as manual counts from
November 2015
2. Test the accuracy of incident
detection alerts recorded by the
system against other sources of
incident data (i.e. CAD system)
Traffic incidents positively identified
through stopped vehicle alerts and
compared with MnDOT CAD logs of
the same events
3. Measure the efficiency of the
system’s operations in all
weather conditions to verify
reliability of system alerts
generated
Vehicle counts and speeds from radar
unit were found to be consistent with
Wavetronix counts and speeds in
November period; Low number of
incident false alarms observed at
beginning and end of the project
4. Test the accuracy of all other
alerts detected against camera
images monitoring traffic along
the corridor and against other
sources of data (i.e. CAD
System)
Alerts generated by 360 Degree
Radar were validated alongside those
events from the MnDOT CAD
system; Low number of incident false
alarms observed at beginning and end
of the project
58
Table 5-2 – Summary of Key Findings from Evaluation of Incident Detection
Capabilities the 360 Degree Radar Information System
Measures Results Notes
Advance Notification Time for
Incidents 10.5 minutes
Represents the average amount of time
that MnDOT RTMC operators would
have had on a total of 11 events
monitored during the project.
Additional Incidents Confirmed
by 360 Degree Radar System 11 events
Represents the number of incidents
detected by the 360 Degree Radar
System that were not detected by
RTMC operators
Incident False Alarm Rates 0.71 / 0.28
Represents the incident false alarm
rates monitored for one week at the
beginning and end of the project.
Table 5-3 – Summary of Operational Needs Addressed by the
360 Degree Radar Information System
General Problems Operational Needs Needs Addressed by System
1. Limited zone
coverage
1. Deploy a traffic detection
system that can monitor multiple
lanes of traffic from one single
location.
System detected vehicles and
incidents in three lanes of WB
and three lanes of EB traffic
from single location
2. Understanding of
latency in incident
detection
2. Gather incident detection
information quickly and accurately
in the project area and provide
information to operational staff for
emergency response
Incidents were detected in a
quick manner by the system.
Alarms were not provided to
RTMC operators, given high
number of alarms created for
stopped traffic.
3. Inability to measure
driver behavior;
individual vehicle
movements
3. Gather traffic congestion
information quickly and accurately
in the project area
Slow moving traffic alarms
were detected, but not found to
useful given traffic congestion
in the area.
4. Inability to perform
multiple real-time data
collection tasks over a
wide area
4. Understand the amount of lane-
weaving occurring within the
project area
Lane weaving was not
measured during the course of
the project.
5. Difficult to gather
performance metrics in a
cost effective manner
5. Gather performance metrics that
can be reported regarding traffic
congestion and incidents in the
project area in a readable and
usable format
Metrics of congestion and
incidents not prepared for use
in performance measure
reporting. Total incidents and
vehicle count summaries were
prepared for the project area.
59
5.2. Considerations for Future Installation
5.2.1. IT Environment Considerations
The 360 Degree Radar unit and associated software provided by Navtech have the
ability to work within existing computer / IT environments through use of existing
server hardware and communications infrastructure in a client-server type of IT
architecture. Users of the software can be established as clients and provided read-
only access to the software showing a live view of the corridor.
While the software can be viewed remotely, Navtech has not provided to training to
TMC operators on the various functions that area available with the Witness software
package. This is primarily because of the complexity involved with the software’s
functions, the firewall and network requirements, and the limited time that TMC
operators have to learn new software packages, in addition to their primary function
of responding to reported 911 incidents and monitoring traffic on a daily basis. Past
clients have simply desired the ability to have alarms generated by the software,
which operators can then act upon by clicking on the alarm, which can bring up a live
camera view of the event detected by the radar.
In the event that assistance is needed in modifying alarms or other features of the
system, Navtech has responded with remote or on-site assistance as necessary. It is
envisioned that a similar arrangement could be established with MnDOT through
agreement either with Navtech directly or through RhiZone, Inc. acting as the local
distributor of the radar unit.
5.2.2. Relocation to Signalized Intersections / DSRC Integration
Through modification to the FCC license previously approved for the radar, the 360
Degree Radar unit could be re-located from its current installation to monitor
signalized intersections given its noted accuracy in generating alerts upon detecting
specific events and in counting vehicles along the I-94 corridor. The central server
that hosts the software package would also need to be installed in the signal cabinet to
provide a connection with the traffic signal controller.
At signalized intersections, the radar and central software could replace in-pavement
loops or other technology currently installed to count vehicles. The radar unit could
be installed at an intersection for a temporary basis to provide counts where no loop
detectors exist to gather a representative sample of vehicle counts for use by traffic
engineers to create or adjust traffic signal timings as needed.
The radar unit and central software could also be used to improve the safety and
efficiency of signal operations under the following types of scenarios:
1. The detection of vehicles at certain distances from the intersection by the radar
and central software could be used to extend signal cycles on main or side street
60
approaches. This could improve the operational efficiency of the intersection
during peak or off-peak travel periods.
2. Stopped vehicles detected by the radar and software on the cross streets of an
intersection could instruct the signal controller to serve those vehicle waiting on
the side street, then return to serving the main approaches of the intersection. This
function could replace the need for in-pavement loops on side street approaches.
3. Vehicles detected by the radar and central software as traveling at high speeds
(greater than “X” mph) as they approach a red light could instruct the signal
controller to hold an all-red phase until the vehicle has either stopped or cleared
the intersection. This could improve the safety of the intersection by reducing
red-light running crashes at the intersection.
4. Pedestrians detected by the radar and central software as crossing the intersection
at a slower pace than allowed to with the given crosswalk time could instruct the
signal controller to hold the “Don’t Walk” until the pedestrian has cleared the
crosswalk. This could improve the safety of the intersection by reducing
pedestrian-vehicle incidents at the intersection.
To enable these types of scenarios, the Navtech server that hosts the software package
would be installed within the signal cabinet and a connection with the signal
controller could be established to allow for the controller to take action under these
scenarios. There are a few approaches that could be taken to enable this connection:
1. The signal controller manufacturer could be the primary lead (by working with
Navtech) to program the signal controller to understand how the inputs from the
radar’s central software can be processed by the signal controller to perform
certain functions.
2. Navtech could be the primary lead (by working with controller manufacturer) on
the development of a program written under an Application Programming
Interface (API), which would allow for the controller to receive inputs from the
radar’s central software and then perform certain functions.
Under either approach, it would be ideal for the traffic signal controller to be in
compliance with the most recent Advanced Traffic Controller (ATC) standard that
guides the development of ATC-model controllers. ATC model controllers operate
under a Linux-based operating system that allows for an API to host the type of
programming that would be required of either the signal controller manufacturer or
Navtech to allow for the radar to instruct the signal controller on how to react to the
types of scenarios noted above.
One consideration that Navtech and signal controller manufacturers could make in the
programming these scenarios is the future implementation of DSRC (Dedicated Short
Range Communications) protocols and the Connected Vehicles program developed
by the Federal Highway Administration. The radar and central software could
supplement this future vehicle-to-intersection (V-2-I) communication by detecting
other vehicles that are not communicating with the controller via DSRC protocols.
The messages sent from the radar central software to the signal controller regarding
61
detection of a moving vehicle can be similar to Basic Safety Messages (BSMs) sent
from Connected Vehicles to the signal controller.
Further investigation would need to be performed by Navtech and the signal
controller manufacturer prior to developing the programs that could enable successful
communication between the central software and the signal controller.
5.2.3. Relocation to Un-Signalized Intersections for ICWS Integration
At un-signalized intersections, the radar and central software could monitor for
vehicles approaching an un-controlled intersection at high speeds where vehicles on
minor side streets are waiting to cross the intersection. MnDOT currently utilizes an
Intersection Conflict Warning System (ICWS) in rural locations to alert vehicles that
may not be able to see vehicles on the main road approaching at high speeds, either
because of the roadway geometrics or a poor line-of-sight with the main roadway.
Similar to signalized intersections, the rural ICWS system utilizes a traffic signal
controller to activate flashing beacons as a warning to drivers when vehicles are
detected at intersection approaches. Since the radar and central software can detect
approaching vehicles from up to 500 meters (about 1,500 feet) away from the
intersection, the radar could successfully be utilized for detection of approaching
vehicles at distances of 1,000 feet from the un-signalized intersection. The central
server that hosts the software package would need to be installed in the nearby signal
cabinet to provide a connection with the existing traffic signal controller.
The radar unit and central software could also be used to improve the safety and
efficiency of ICWS operations under the following types of scenarios:
1. The detection of vehicles at 1,000 feet on the main approach from the intersection
by the radar and central software could be used to activate flashing beacons that
issue warnings to motorists on the cross street. This would operate similar to how
existing in-pavement loops activate these flashing beacons.
2. The detection of vehicles entering the intersection from the side street by the radar
and central software could be used to activate flashing beacons that issue
warnings to motorists on the main approach. This would operate similar to how
existing in-pavement loops activate these flashing beacons.
3. Pedestrians detected by the radar and central software as crossing the un-
signalized intersection could also trigger flashing beacons as a warning to
motorists on the main approach. This could improve the safety of the intersection
by reducing pedestrian-vehicle incidents at the intersection, and provides
additional functionality to the ICWS that is not available with the in-pavement
loops currently installed.
4. The radar and central software could also be used to monitor how vehicles are
currently reacting to the warnings presented to them on roadside signage that
illuminates as part of the ICWS installation. Recordings of vehicle speeds on
main approaches when vehicles enter an intersection could be established to
understand how vehicle speeds change with the flashing beacons activated at
62
various times of the day. This type of analysis could lead to potential adjustments
in how the ICWS is installed to improve the overall safety of the ICWS at the
intersection.
To enable these types of scenarios, the Navtech server that hosts the software package
would be installed within the signal cabinet and a connection with the signal
controller could be established to allow for the controller to take action under these
scenarios. There are a few approaches that could be taken to enable this connection:
1. The signal controller manufacturer could be the primary lead (by working with
Navtech) to program the signal controller to understand how the inputs from the
radar’s central software can be processed by the signal controller to activate the
flashing beacons.
2. Navtech could be the primary lead (by working with controller manufacturer) on
the development of a program written under an Application Programming
Interface (API), which would allow for the controller to receive inputs from the
radar’s central software and then perform certain functions.
Current installations of the ICWS utilize non-ATC model signal controllers since the
advanced functionalities on an ATC are not necessary for an un-signalized
intersection. It is likely that the signal controller manufacturer could program the
signal controller to respond to inputs from the central software by activating the
appropriate flashing beacons. Further investigation would likely be needed by
Navtech and the signal controller manufacturer prior to implementing this type of
application.
The radar unit could also be installed on a temporary basis to gather and record data
for a one week period of time, prior to being re-located to another ICWS installation
in other areas of the state as well. This work would involve installing the radar unit
and central software onto a mobile trailer where it could be moved to various
locations to collect data on incidents in specific areas, or perform incident detection
for a limited period of time as needed.
5.2.4. Use of Alarms with ATM Signage on Roadway
The alarms generated from the 360 Degree Radar Information System could be used
in multiple other manners by MnDOT or traffic operations staff. One example of the
use of slow-moving vehicle alarms could be to trigger messages to appear on
Dynamic Message Signs (DMS) that are installed as part of the Active Traffic
Management (ATM) systems installed on I-35W and I-94.
Currently, speeds displayed on the DMS that alert traffic to a recommended speed
based on the speed of traffic detected approximately one half-mile upstream from the
DMS. A slow-moving vehicle alarm from the 360 Degree Radar unit could trigger a
slower recommended speed, such as 15 or 20 MPH, that would reduce the potential
for crashes from traffic attempting to stop suddenly at unsafe speeds.
63
Other messages could be displayed on these DMS reading “Slow / Traffic / Ahead”
utilizing three lines of text on the DMS that are used as part of the ATM system. This
type of message is currently displayed on those DMS as heavy congestion is detected,
but these messages could be automated through use of the 360 Degree Radar unit.
This could require the development of an XML (Extensible Markup Language) feed
between the 360 Degree Radar central software that detects slow-moving and stopped
traffic alarms and the MnDOT IRIS system utilized to activate messages on DMS.
The XML feed would enable the alarms generated by the 360 Degree Radar central
software to allow for the activation of the messages on the DMS that are monitored
and controlled by MnDOT. Navtech Radar has developed similar types of XML
feeds for alarms for other clients and could do so for MnDOT after further discussion
of how the alarms could activate DMS messaging.
Further testing of the system and its ability to accurately report vehicle speeds during
cold weather periods would likely be required prior to implementing an XML feed as
described above. As noted earlier in this report, due to changes that were made to
system parameters within the software on Jan. 15th, the accuracy of speed
measurements in the winter period of time negatively impacted the ability of the
system to accurately measure speeds in this period of time.
5.3. Recommendations for Future Installation
This section contains next steps for MnDOT to consider taking with the 360 Degree Radar
Information System. Further discussions with the project team can take place in subsequent
meetings to determine how these steps can be taken to help improve the overall safety of the
transportation system.
5.3.1. Integrate into MnDOT RTMC Operations
This step could be taken to gather feedback from RTMC operators about the overall
operation and reliability of the system to assist them in detecting incidents in the
project area of I-94 between 3rd Avenue and the MN-65 overpasses.
It should be noted that MnDOT did allow for a limited trial period of the system
within the MnDOT RTMC for two weeks in late May and early June. A MnDOT
supervisor monitoring the system noted that many stopped vehicle alarms were
generated during periods of heavy traffic congestion on the corridor. While the
alarms were accurate for stopped traffic, the value of alarms for incident detection
was very limited given the frequency of traffic congestion in the area. As a lesson
learned for this traffic environment, Navtech proposed a modification to the software
that could limit the number of alarms generated in heavy traffic congestion.
RhiZone has also recommended an adjustment to the tilt of the radar to further
improve the detection of traffic incidents on the corridor. Adjustments to the central
software would then take approximately five days to account for the change in the
radar tilt toward the roadway. An XML feed could then be established on the project
server to send alerts directly to RTMC operators.
64
Navtech Radar is currently working on a software upgrade to better detect traffic
incidents when there is a high amount of congestion on the corridor. This upgrade
could be provided prior to alerting RTMC operators about incidents in the project
area.
5.3.2. Integrate System into a Signalized Intersection
A second step that could be taken in future years would be to move the 360 Degree
Radar Information System to a signalized intersection where it could perform the
simultaneous tasks of vehicle count / speed measurements, as well as incident
detection. This project could involve Navtech Radar working closely with a signal
controller manufacturer to determine how inputs from the radar in the field could feed
the signal controller (or the controller’s central software package) to actuate signal
operations in the field. Purpose of this project would be to see how well the radar
could replace existing in-pavement loops currently used for vehicle detection.
5.3.3. Integrate System onto a Mobile Trailer
A third step that could be taken in future years would be to move the 360 Degree
Radar Information System onto a mobile trailer where it could be moved to various
locations to collect data on incidents in specific areas, or perform incident detection
for a limited period of time as needed.
65
Appendix A – System Component Data Sheets
Pages 1-2 – 360 Degree Radar Specification
Pages 3-5 – Proposed Mounting Bracket and Detail for 360 Degree Radar
Pages 6-9 – Radar Power Supply Unit
Pages 10-13 -- Proposed CCTV Camera Brackets
Pages 14-17 – Proposed Axis 1614-E Camera Datasheet
Pages 18-19 – Proposed CCTV Camera Brackets
Pages 20-22 – Proposed Software Application for 360 Degree Radar System
Page 23 – Network Video Recorder Datasheet