INCIDENT DETECTION AND ARTIFICIAL NEURAL NETWORKS
In this paper, the Back Propagation Technique with Conjugate Gradient Minimization was used on simulated traffic data. The simulated traffic data was obtained from a calibrated INTEGRATION model of the Toronto section of the Highway 401 - Freeway Traffic Management System (FTMS). The data consisted of volume, occupancy, and speed information. Supervised learning was used with different data set input patterns. Independent data sets were used for training and testing. The sensitivity to both time and spatial data was evaluated. Using this method, it was found that the detection probability was high and there was a low probability of false alarm.
- First in a series on Management of Traffic Systems conducted by Alberto J. Santiago and Sam Yagar on behalf of the Federal Highway Administration (FHWA).
Washington, DC United States 20590
- Wiederholt, L
- Okunieff, P
- Wang, J
- Large Urban Systems. Proceedings of the Advanced Traffic Management Conference
- Location: St. Petersburg, Florida
- Date: 1993-10-3 to 1993-10-8
- Publication Date: 1993-10
- Features: Figures; References; Tables;
- Pagination: v.p.
- TRT Terms: Freeways; Highway traffic; Incident detection; Traffic simulation; Traffic volume
- Geographic Terms: Toronto (Canada)
- Old TRIS Terms: Freeway traffic
- Subject Areas: Highways; Operations and Traffic Management; I73: Traffic Control;
- Accession Number: 00662944
- Record Type: Publication
- Files: TRIS, USDOT
- Created Date: Jul 7 1994 12:00AM