On Using Battery State for Medium On Using Battery State for Medium Access Control in Ad hoc Wireless Access Control in Ad hoc Wireless

Networks - “BAMAC”Networks - “BAMAC”

S. Jayashree, B. S. Manoj, C. Siva Ram MurthyS. Jayashree, B. S. Manoj, C. Siva Ram Murthy

Department of Computer Science and Engineering,Department of Computer Science and Engineering,

Indian Institute of Technology Madras,Indian Institute of Technology Madras,

India.India.

Energy Management in Ad hoc Energy Management in Ad hoc Wireless NetworksWireless Networks

Energy management:Energy management: Process of managing energy resources by Process of managing energy resources by means of controlling the battery discharge, adjusting transmission means of controlling the battery discharge, adjusting transmission power, and scheduling power sources, so as to increase the lifetime of power, and scheduling power sources, so as to increase the lifetime of an Ad hoc wireless networkan Ad hoc wireless network

Why energy management?Why energy management? Fig

Limited energy reserveLimited energy reserve

Difficulties in replacing the batteriesDifficulties in replacing the batteries

Lack of central coordinationLack of central coordination – – relay nodes?relay nodes?

Constraints on the batteryConstraints on the battery – – weight-capacityweight-capacity relationshiprelationship

Selection of optimal transmission powerSelection of optimal transmission power

Channel utilization for CDMA based systemChannel utilization for CDMA based system

Battery-related TermsBattery-related Terms

What is a battery ?What is a battery ? Characterizing batteriesCharacterizing batteries

• Voltages (open circuit, operating, and cut-off)Voltages (open circuit, operating, and cut-off)

• Capacities - Rate capacity effect and Recovery capacity Capacities - Rate capacity effect and Recovery capacity effecteffect

o Theoretical (T)Theoretical (T)

o Nominal/standard (N) Nominal/standard (N)

o ActualActual

BAMAC(BAMAC(kk) MAC Protocol) MAC Protocol Inspiration!Inspiration! Battery-aware MAC protocol (BAMAC(Battery-aware MAC protocol (BAMAC(kk))))

• Main Idea – MAC + Battery informationMain Idea – MAC + Battery information• Exploiting recovery state of the batteriesExploiting recovery state of the batteries• Uniform discharge of the batteriesUniform discharge of the batteries• Nodes are scheduled in a near round-robin mannerNodes are scheduled in a near round-robin manner• MAC protocol – “The higher the remaining battery capacity, the MAC protocol – “The higher the remaining battery capacity, the

lower the back-off period.”lower the back-off period.” Discrete-time Markov chain analysis of battery lifetimeDiscrete-time Markov chain analysis of battery lifetime Comparative study of IEEE 802.11 and DWOP with Comparative study of IEEE 802.11 and DWOP with

BAMAC and Simulation Vs. Theoretical AnalysisBAMAC and Simulation Vs. Theoretical Analysis Analysis of the factor ‘Analysis of the factor ‘kk’’

Illustration of Battery TableIllustration of Battery Table

Back-off CalculationBack-off Calculation

CwCwminmin – Minimum size of the contention window– Minimum size of the contention window

rankrank – – Position of that entry in the batteryPosition of that entry in the battery TTSIFSSIFS+T+TDIFSDIFS – SIFS and DIFS duration as in IEEE 802.11 – SIFS and DIFS duration as in IEEE 802.11

TTtt – Successful packet transmission time, including RTS- – Successful packet transmission time, including RTS-

CTS-Data-ACKCTS-Data-ACK x x – Number of transmission attempts made for a packet– Number of transmission attempts made for a packet Uniform(Uniform(i,ji,j) – random number distributed uniformly ) – random number distributed uniformly

between between i i and and jj n - n - Number of neighborsNumber of neighbors

Illustration of BAMACIllustration of BAMAC

T- Theoretical capacity

N- Nominal capacity

t- Time of transmission

Modeling the Batteries Using Modeling the Batteries Using Discrete-Time Markov Chain With Discrete-Time Markov Chain With

Probabilistic RecoveryProbabilistic RecoveryNotations used:Notations used:

TT - Initial theoretical capacity ; - Initial theoretical capacity ; NN - Initial nominal capacity- Initial nominal capacity TTii - Theoretical capacity at time unit - Theoretical capacity at time unit I; I; NNii - Nominal capacity at time unit - Nominal capacity at time unit i i

TxTx -- Transmission state ;Transmission state ; RxRx -- Recovery stateRecovery state RRNi,TiNi,Ti - Probability to recover one charge unit - Probability to recover one charge unit

Illustration of battery discharge Illustration of battery discharge of nodes using BAMAC(k)of nodes using BAMAC(k)

Average discharge time = Average discharge time = k*Tk*Ttt

Average recovery time = Average recovery time = n*k*Tn*k*Ttt

n*kn*kk

PM

Final Markov Model Representing Final Markov Model Representing Battery BehaviorBattery Behavior

jiQQ ,

Steps to Calculate Time Duration of the Markov Model to Remain in Steps to Calculate Time Duration of the Markov Model to Remain in Transient StatesTransient States

Performance AnalysisPerformance Analysis

GloMoSimGloMoSim Simulation parametersSimulation parameters

Transmission – Transmission – 22; Reception –; Reception – 1 1 ; idle – ; idle – 0 0; Listening -; Listening - 0 0 A small battery to power up during idle modeA small battery to power up during idle mode Existence of SBS – state of the batteryExistence of SBS – state of the battery

Simulation areaSimulation area 2000m x 2000m2000m x 2000m

Number of nodesNumber of nodes 10-4010-40

Transmission powerTransmission power 12dB12dB

Channel bandwidthChannel bandwidth 2Mbps2Mbps

Routing protocolRouting protocol DSRDSR

Path loss modelPath loss model Two-RayTwo-Ray

Battery ParametersBattery Parameters T=2000; N=250; g=0.05T=2000; N=250; g=0.05

Results for BAMAC(Results for BAMAC(kk))Number of packets transmitted Remaining battery charge

Infinite theoretical capacity

……ContdContd

Average throughput Standard deviation of throughput

Remaining nominal capacity Nominal capacity spent

……ContdContd

Packets transmitted for 1>k>20 Packets transmitted for 1>k>250

Simulation results

Theoretical results

……ContdContd

Throughput for 1>k>20 Throughput for 1>k>250

Conclusion and Future PerspectiveConclusion and Future Perspective What is our contribution?What is our contribution? BAMAC(BAMAC(kk) – Battery Aware MAC protocol) – Battery Aware MAC protocol Modeling of batteries using Discrete-time Markov chain analysis: Battery Modeling of batteries using Discrete-time Markov chain analysis: Battery

Lifetime CalculationLifetime Calculation Analysis of the factor “Analysis of the factor “k”k” Performance AnalysisPerformance Analysis Future perspectiveFuture perspective Finding optimalFinding optimal K K ValueValue Relaxing the basic assumptionsRelaxing the basic assumptions

Sleep mode instead of idle mode!Sleep mode instead of idle mode! Absence of an additional small batteryAbsence of an additional small battery

Heterogeneous battery technologiesHeterogeneous battery technologies Presence of real-time data trafficPresence of real-time data traffic Generalized Markov modelGeneralized Markov model Develop a generalized tool for calculating battery life time for all MAC Develop a generalized tool for calculating battery life time for all MAC

protocolsprotocols

THANK YOUTHANK YOU

IIT MadrasIIT Madras

IBM-IRL IndiaIBM-IRL India

DST Delhi, IndiaDST Delhi, India

Increasing Gap Between Power Increasing Gap Between Power Requirement And AvailabilityRequirement And Availability

K. Lahiri, A. Ragunathan, S. Dey, and D. Panigrahi, “ Battery-DrivenSystem K. Lahiri, A. Ragunathan, S. Dey, and D. Panigrahi, “ Battery-DrivenSystem Design: A New Frontier in low-Power Design,” Design: A New Frontier in low-Power Design,” Proceedings of ASP-DAC/VLSI Proceedings of ASP-DAC/VLSI Design 2002.Design 2002.

back