Traffic Scheduling For Energy Sustainable Vehicular Infrastructure

Abdulla A. H., Ghada H. B., Terence D. T., Amire A. S. and Dongmei Z.

IEEE Globecom 2010 proceeding

Presented by 劉美妙

OutlineIntroductionSystem model and optimal energy boundNearest fastest set (NFS) schedulerPerformance evaluationConclusions

IntroductionVehicular network infrastructure will eventually evolve into a

platform which will permit an even larger variety of mobile applications.

In many highway locations, setting up this type of infrastructure is difficult due to the unavailability or excess expense of wired electrical power.

An alternative to wired power connections is to operate some of the APs using an energy sustainable source such as solar or wind power.

Since multiple vehicles may be present in the AP coverage area, the question arises as to the order with which vehicles should be served.

As the energy available for the AP is limited, it is important to design energy efficient scheduling algorithms.

System model and optimal energy boundSingle roadside AP

Serving vehicles passing by in one directionSingle radio transceiverConnect to one vehicle at any given time for a time slot, Δt

Pdoze: power consumption in Doze modethe AP’s radio is placed into a Doze mode when traffic is not

being processed

Pt : AP transmission power at time ttransmit to vehicles using as small a transmit power as possiblesubject to satisfying the communication requests issued by the

passing vehicles

Minimize the total energyA standard exponential distance-dependent path loss

model is used.

power consumption in Doze mode

Energy used by AP to serve the demands of the vehicles

Nearest fastest set (NFS) schedulerNearest Fastest Set (NFS) Scheduler that uses received

knowledge of vehicle locations and velocities over a given time window

NFS: schedule vehicle requests when they are as close as possible to the AP.same distance: determined by velocity of the vehicles

In a given time slot, NFS decides the order with which vehicles will be served for a short duration of time

NFS consists of two execution phasesPreparation of candidate communication opportunities.Scheduling phase

Preparation of candidate communication opportunities.

NCLNv × tws=

Vehicle v1, R1=4, ws=5Select nearest locations

NFS Scheduler-phase I

t1 tws…

Time window

V1

VNv

… …

t1 t2 t3 t4 t5

v1

v2

v3

v4

AP

Dv1,t5Dv1,t3Dv1,t2Dv1,t1 Dv1,t4

Distance: Dv1, t3 < Dv1, t4 < Dv1, t5 < Dv1, t2 < Dv1, t1

1 1 11

NFS Scheduler-phase IIScheduling phaseresolves contention between vehicle requestsIn time slot t,If , then , else if

Compute weight:

t

v1 1 1 1

v2 1

v3 1 1

1

1

1

AP

t

The simulation time for each run is 600 time steps and the number of vehicles is 48.

The window size for the NFS and NO Schedulers is 10 time steps.

NFS scheduler performance for different traffic density

Performance evaluation

NFS scheduler performance for different σ

NFS schedule performance for different α

Density: 8 Vehicles per KM (125m apart)Average speed: 72KM/h

Performance evaluation

Different demand Different traffic density

Standard deviation σ = 4,Propagation path loss exponent α = 2,Density: 4 Vehicles per KM (125m apart)

Demand level R = 6, Average speed: 108KM/h

Compare with other scheduler

Different σ Different α20 times

5 times

Compare with other scheduler (cont.)

ConclusionsIn this paper, we have considered the problem of

satisfying vehicular communication requirements while minimizing the energy needed by the roadside access point.

proposed a Nearest Fastest Set (NFS) scheduler that uses vehicle location and velocity inputs to perform causal scheduling.

since multiple vehicles may be present in the AP coverage area, the question arises as to the order with which vehicles should be served.phraseas to something;   as regards something關於 ;至於

AP

Dv1,tDv2,t

Dv3,t