METRANS

10-05 End-to-End Performance in Vehicular Networks With an Emphasis on Safety and Security Applications

Project Number

10-05

Project Summary

End-to-End Performance in Vehicular Networks With an Emphasis on Safety and Security Applications

Project Status

Complete

Project Report

Year

2010

Topic Area

Safety, Security, & Vulnerability

P.I. Name & Address

Associate Professor, Ming Hsieh Department of Electrical Engineering; USC Viterbi School of Engineering
University of Southern California
3710 McClintock Ave.
Ronald Tutor Hall (RTH) 409
Los Angeles, CA 90089-2562
United States
kpsounis@usc.edu

Project Objective:

The objective of this current proposal is to address the challenges related to end-to-end performance. Specifically, the PI plans to work on the following central topics: (i) Congestion control: To provide good end-to-end performance, networks employ some form of congestion control to throttle data rates when the network cannot accommodate them. Traditional congestion control schemes rely on "stream-oriented" transport protocols like TCP. But, stream-oriented transport is impossible in the context of intermittently connected networks. With this in mind, the PI will adopt a "message-oriented" transport approach, and design congestion control schemes which work efficiently on top of such transport protocols. (ii) Interoperability with connected networks: Roadside stations are envisioned to be connected to Internet-like, fully connected networks. The fundamental differences between a VANET and a traditional connected network require roadside stations to run specialized schemes which make the networks interoperable. The PI plans to investigate such schemes to enable end-to-end connectivity when one of the nodes resides in a VANET (e.g. moving vehicle) and the other in a connected network (e.g. a police station). (iii) Network characterization: Finally, a very important factor affecting end-to-end performance is the inherent homogeneity in VANETs. Specifically, since the level of connectivity may change fast in space and time, the PI plans to design automated distributed mechanisms that allow nodes to characterize on the fly how connected the network is, and select appropriate protocols given the conditions. In addition to designing schemes that address the challenges above, the PI plans to run simulations using real-world maps and traces, as well as experiments in collaboration with the USC Transportation Office.

Also to design highly efficient vehicular ad hoc networks (VANETs), that are tailored to supporting real-world safety- and security-related applications. Realistic simulations that use real roadmaps and mobility traces, and formal analysis tht takes real-world limitations into account, will be used to design the proposed algorithms and access their performance. Specific safety applications will be used as showcases to establish the applicability and efficiency of the proposed schemes in timely delivering of real-world applications.

Task Descriptions

Milestones, Dates:
2009

Total Budget:
$90,000

Student Involvement:
One graduate student at 50% effort, 9.36 months

Relationship to Other Research Projects:
Related to 08-12; part of the mobility focus area

Technology Transfer Activities:
Project report will be posted soon

Potential Benefits of the Project:
Improving the Safety of Drivers, Passenegers, and Pedestrians

TRB Keywords:
VANETs, ad hoc, Psounis

Primary Subject:
Abs.3 To use a new approach of communication that is tailored to the needs of vehicular ad hoc networks and is termed by the PI as "mobility-assisted communication."