In this paper, an intelligent-self learning dynamic optimal contraflow lane control system developed for the George Massey Tunnel in southern Greater Vancouver is introduced. A program was developed to permit the accurate estimation of the real-time traf€ic demands. On-line traffic data are sorted by a fuzzy modeling algorithm to identlfy the best matching pattern. A self-learning mechanism is utilized to modify the predicted demand incrementally. An optimization algorithm is developed for on-line calculation of the optimal contraflow schedule based on the predicted demand. The total delay of both traffic approaches is minimized.This intelligent traffic contraflow lane control system has been implemented into a generic form of program package written in C. It can be applied to other similar systems where traffic contraflow lane controls are required. 1 0-7803-1235-X/93/$3.00 0 1993 IEEE

No abstract

No abstract