Date of Award

Spring 2009

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Computing Sciences - (Ph.D.)

Department

Computer Science

First Advisor

Cristian Borcea

Second Advisor

Marvin K. Nakayama

Third Advisor

Guiling Wang

Fourth Advisor

Nirwan Ansari

Fifth Advisor

Teunis J. Ott

Sixth Advisor

Constantin Serban

Abstract

Vehicular ad hoc networks (VANETs) can provide scalable and cost-effective solutions for applications such as traffic safety, dynamic route planning, and context-aware advertisement using short-range wireless communication. To function properly, these applications require efficient routing protocols. However, existing mobile ad hoc network routing and forwarding approaches have limited performance in VANETs. This dissertation shows that routing protocols which account for VANET-specific characteristics in their designs, such as high density and constrained mobility, can provide good performance for a large spectrum of applications.

This work proposes a novel class of routing protocols as well as three forwarding optimizations for VANETs. The Road-Based using Vehicular Traffic (RBVT) routing is a novel class of routing protocols for VANETs. RBVT protocols leverage real-time vehicular traffic information to create stable road-based paths consisting of successions of road intersections that have, with high probability, network connectivity among them. Evaluations of RBVT protocols working in conjunction with geographical forwarding show delivery rate increases as much as 40% and delay decreases as much as 85% when compared with existing protocols.

Three optimizations are proposed to increase forwarding performance. First, one- hop geographical forwarding is improved using a distributed receiver-based election of next hops, which leads to as much as 3 times higher delivery rates in highly congested networks. Second, theoretical analysis and simulation results demonstrate that the delay in highly congested networks can be reduced by half by switching from traditional FIFO with Taildrop queuing to LIFO with Frontdrop queuing. Third, nodes can determine suitable times to transmit data across RBVT paths or proactively replace routes before they break using analytical models that accurately predict the expected road-based path durations in VANETs.

Share

COinS