DOI

10.5703/1288284313347

Abstract

This study develops an Internet-based remote on-line traffic control architecture for route guidance in large-scale traffic systems equipped with advanced information systems and sensors. It also proposes to use the Beowulf Cluster paradigm as an economical, flexible, and customizable computing architecture to generate supercomputing capabilities within the control architecture. The Beowulf Cluster provides the enabling environment to execute and coordinate the activities of the various modules that address real-time network route guidance, data transmission, calibration and fault tolerance. In the context of large-scale traffic systems, a Beowulf Cluster can be configured in centralized as well as decentralized on-line control architectures with equal ease. Thereby, it enables individual traffic operators with smaller operational scope (such as local traffic agencies) to install mini Beowulf Clusters at their locations or allows several of them to operate remotely using a centrally located large-scale Cluster. To enable automation and the seamless operation of the on-line architecture, a Fourier transform-based fault tolerant framework is introduced that can detect data faults due to malfunctioning detectors and predict the likely actual data on-line.

Report Number

FHWA/IN/JTRP-2001/07

Keywords

Internet-based On-line Traffic Control Architectures, Beowulf Clusters, Real-time Route Guidance Deployment, Fault Tolerance, Automatic Detection of Sensor Malfunctions, Real-time Traffic Data Communication and Storage, SPR-2399

SPR Number

2399

Project Number

C-36-75M

File Number

8-9-13

Performing Organization

Joint Transportation Research Program

Publisher Place

West Lafayette, IN

Date of this Version

2002

Share

COinS