Browse Prior Art Database

A System for Improved Traffic Management/prediction

IP.com Disclosure Number: IPCOM000029728D
Original Publication Date: 2004-Jul-09
Included in the Prior Art Database: 2004-Jul-09
Document File: 2 page(s) / 33K

Publishing Venue

IBM

Abstract

Traffic congestion is a common problem in the automotive highway system. Dynamic conditions such as traffic load, road construction and accidents introduce problems for static route determination in travel planning. Many on-line mapping systems (such as MapQuest) provide the ability to obtain shortest or fastest path travel directions between locations, using static information. This disclosure extends road-based sensor networks to include GPS enabled automobiles (thereby supporting all roads, rather than just the cities/interstates with sensor networks) and to apply known flow/congestion control algorithms from computer networking to the traffic problem.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 2

A System for Improved Traffic Management /prediction

Traffic congestion is a common problem in the automotive highway system. Dynamic conditions such as traffic load, road construction and accidents introduce problems for static route determination in travel planning. Many on-line mapping systems (such as MapQuest) provide the ability to obtain shortest or fastest path travel directions between locations, using static information.

Systems to record traffic accidents are common in many countries. Tracking traffic load using sensor networks has recently been introduced in the United States. These systems provide necessary input for determining the number of vehicles on particular roads/highways. An alternative system could track automobiles via GPS devices on the cars. In both approaches, a straightforward approach to traffic flow prediction is to make assumptions about the flow network. Prediction systems using only this technology (as in http://www.msnbc.msn.com/Default.aspx?id=3158259&p1=0) suffer from a number of shortcomings. First, they do not monitor the number of vehicles considering a re-route or actually re-routing to alternate routes and create a "thundering herd" problem where many vehicles head to what was once the quickest route. (see load balancing below) Second, more sophisticated routing techniques exist which can be applied to traffic situations to reduce congestion and provide improved flow control.

This disclosure extends road-based sensor networks to include GPS enabled automobiles (thereby supporting all roads, rather than just the cities/interstates with sensor networks) and to apply known flow/congestion control algorithms from computer networking to the traffic problem. Specifically, we propose using the portion of High-Performance Routing (HPR) known as Adaptive Rate Based flow control (ARB) for dynamically determining route al...