Design of On-line Simulation System for Signal Control of Urban Intersections Based on Visual Sensing Technology
Juanjuan Wang,Yanan Wang,Hongfang Zhou
DOI: https://doi.org/10.1088/1742-6596/1982/1/012059
2021-07-01
Journal of Physics: Conference Series
Abstract:Abstract Intersection area is the area with the most serious conflicts and accidents in the whole urban traffic system. Therefore, it is of great significance to combine modern computer technology, sensing technology and communication technology to intelligently control intersections, improve the efficiency of traffic signal control and reduce the delay time in conflict areas. Starting with video detection and using deep learning in the detection module, a video detection mechanism based on restricted Boltzmann machine network is proposed. Based on C # programming language, the intersection traffic signal simulation module is designed through COM interface provided by traffic simulation software VISSIM, which realizes road network import, parameter setting, signal timing change, simulation operation and simulation result output. Numerical experiments show that fuzzy control method with self-correcting factor is superior to timing control method and classical fuzzy control method in average delay time of vehicles.