A Mobile Infrastructure Based VANET Routing Protocol in the Urban Environment School of Electronics Engineering and Computer Science, PKU, Beijing, China IEEE CMC 2010 Jie LuoXinxing GuTong ZhaoWei Yan

Outline  Introduction  Related Work  Urban VANETs Analysis  MIBR Design Routing Forwarding  Performance Evaluation  Conclusion

Introduction  Vehicular ad hoc networks have recently received considerable attention  The VANET provides both Roadside-to-Vehicle communication Inter-Vehicle communication (IVC)  Works like a MANET with its own unique characteristics

Introduction  Key Challenge of Urban VANET The existence of frequent network disconnection is one of the key challenges for routing protocols for urban VANETs

Introduction  Goal Analyze the features of urban VANET Propose a routing protocol MIBR to improve the connectivity of the network by taking advantage of urban characteristics

Related Work  GPSR GPSR selects the node that is the closest to the destination among the neighboring nodes S A B D GPSR will choose B, because B is closer to D than A. Greedy Perimeter Stateless Routing for Wireless Networks. MobiCOM 2000

Related Work  RAR A hybrid routing protocol with vehicles and RSU (Road Side Unit) Roads are divided into sectors by RSU The drawback of this protocol is the requirement and distribution of static node or RSU

Urban VANETs Analysis  Vehicle movements are constrained by roads in urban environment The routing in urban VANET should be a sequence of road segments The decision to choose which road segment near the junction for forwarding is critical

 Traffic lights have great influence on the vehicle movement Urban VANETs Analysis Vehicles are moving like a cluster

Urban VANETs Analysis  Vehicles have at least two different types in the urban environment Ordinary cars  Less than bus (80% of ordinary cars in Beijing) Buses  More than ordinary cars (20% of buses in Beijing)  Larger and more powerful  Can carry better wireless equipment with a larger transmission range than ordinary cars

MIBR Design  Assumptions Each vehicle knows its location through GPS Each vehicle has a digital street map including bus line information Source node can get the information of destination location

MIBR Design  Assumptions Each bus has two wireless interfaces working on different channels  R1 and R2 Ordinary car has only one interface  R1 R1 R2 R1 Transmission range between buses Transmission range between cars and between cars and bus

MIBR Design  Overview Routing  Selecting an optimal route which consists of a sequence of road segments with the best estimated transmission quality Forwarding  Efficiently forwarding packets hop-by-hop through each road segment in the selected route

MIBR Design  Routing Calculate hop count number of each road segment by the density of buses on the road segment  MIBR prefers road segment with less hop count number Dijkstra algorithm would be used to select a shortest route with the minimal expected hop count The next road segment would be chosen when packet is near a junction

MIBR Design A B

A B

A B

 Forwarding Bus first strategy  Send the packet to the node on the next road segment when it is near the junction  The decision of next hop near the junction is critical  Buses have higher priorities to be the next hop because of the transmission range between buses is larger

MIBR Design  Forwarding If there are any buses on the next road segment, choose the one which is the closest to the junction after the next junction. Otherwise, choose the ordinary car If there are no vehicles on the next road segment, and packet is now on a bus or on an ordinary car, choose a bus which is closest to the next junction. Otherwise, choose the ordinary car

MIBR Design  Forwarding If there are no better suitable forwarding nodes, drop the packet

Performance Evaluation  Simulation Model Ns-2 simulator Simulated area is based on Southern Beijing with a 1700m*1000m size in real world The vehicle movement trace is generated by VanetMobiSim

Performance Evaluation Parameter Value R1150m R2300m Bandwidth for both channel2Mb Beacon interval1.0s Vehicle velocity0-30m/s Number of nodes The bus percent20% Packet size512bytes Simulation time600s Simulation Parameters Table

Performance Evaluation  The data delivery ratio in different network density

Performance Evaluation  Throughput of networks with 200 nodes

Conclusion  The proposed protocol is a geographical routing using the map topology and the bus line information  The algorithmic complexity of MIBR is low, and the deployment is easy because no static nodes or RSUs are needed in MIBR

THANK YOU