Fuzzy Neural Network Control Technique and Its Application in A Complex Intersection

GJ Shen,YX Sun
DOI: https://doi.org/10.1109/icnnb.2005.1614781
2005-01-01
Abstract:For a lack of complex intersection effective control method, a novel single intersection intelligent control technique was presented. Signal lamp setting, phase division, pedestrian press-button response, signal lamp advance-off and late-on were discussed in detail. Being a function of traffic flow saturation degree the cycle time was optimized by one fuzzy neural network, and the variable splits were adjusted based on historical and real-time traffic information. The goal was to make the intersection expedite, vehicles stop rate shortest and average - vehicle delay time less. However, the crossing-crosswalk safety of pedestrian had to be assured firstly. This control technique has been used successfully in the intersection of No. 9 road and No. 12 road in Xiasha economic technical development area Hangzhou, China, and the application results have shown to be very good.
What problem does this paper attempt to address?