Video-based Technology Of Optimizied Signal Control For Pedestrian Crossing

In urban road sections and intersections,pedestrian crossings are often accompanied by the competition for time and space resources with other transports such as motor vehicles and non-motor vehicles.For many years,in urban transportation planning of China,the thought of emphasizing motor vehicles and ignoring non-motor vehicles,including walking and other modes of transport,has been deeply rooted.In recent years,some cities have paid more attention to pedestrian traffic.Some design agencies are starting to tilt more resources for pedestrians.According to the requirements of relevant laws of China,the motor vehicles should yield to the vulnerable pedestrians.But the situation is still pessimistic,the safety of pedestrians crossing on the street is not guaranteed.Large amount of disorder walkers and vehicles conflicts together that is caused by the uncivilized behaviors.These include scrambling for roads,jaywalk,nonobservance of traffic lights and other behaviors of walkers.Therefore,to increase operation efficiency and reduce waiting time of pedestrians,it is necessary to establish a pedestrian crossing signal optimization control method for the complex traffic environment in China,which based on the comprehensive study of pedestrian traffic characteristics and the operation of motor vehicles.The design implements a software system that integrates the above functions.This paper mainly includes two aspects: control method research and system implementation.Control method research is the theoretical basis of system implementation.On the other hand,system implementation is the practical application of control method.The experimental results show that the theoretical research is correct.The research of control method mainly includes three parts: Pedestrian Count Estimation Method;pedestrian traffic characteristics analysis and modeling and study on pedestrian optimal control.To reduce the possible external disturbance,video images containing different traffic conditions are processed.After this,based on using the causality between image features,a Bayesian network is established to infer the number of pedestrians,and finally a more accurate detection value of pedestrians is obtained.Then it mainly studies the pedestrian traffic characteristics,including the speed characteristics,psychological characteristics(waiting time),space-time characteristics of pedestrian trajectories,the density characteristics of the crowd and the retardation effect of two-way pedestrians crossing the street.At the same time,based on the previous pedestrian detection values,the relationship model between pedestrian crossing time and pedestrian number is established,which provides a theoretical basis for the study of signal optimization control method.Based on the phase model of pedestrian crossing characteristics,two pedestrian crossing control systems,timing control method for road section and coordinated control method for section to intersections,are proposed for different traffic environments,with the corresponding case analysis of actual traffic environment of intersections and sections.From the aspect of time resources,according to the different traffic conditions of adjacent intersections,coordinated control method of pedestrian phase is carried out.Combining with pedestrian detection algorithm,an optimization model of pedestrian crossing is established to determine the length of green wave of upstream and downstream coordinated phase and pedestrian crossing phase.From the aspect of space resources,in order to ensure the safety of pedestrians passing through the wide section,the pedestrian safety island is set up on the long pedestrian crossing to achieve twice crossing.The coordinated optimal control of the first crosswalk and adjacent intersections achieve the minimum delay.By comparing different pedestrian crossing models with VISSIM simulation software and COM interface,the effectiveness of the proposed control method is proved.In the aspect of system implementation,the pedestrian detection equipment is integrated through software and hardware design.The software part is embedded software for the algorithm of pedestrian video detection equipment and pedestrian crossing time algorithm.The hardware part includes six parts: image processing system,video acquisition system,video output system,communication system,chip memory chip and power supply design.All parts have specific chip selection and circuit design.The major tentative conclusions of this study can be summarized as follows:(1)a video detection device which can accurately estimate the number of pedestrians crossing the street is designed.(2)Based on the estimated number of pedestrians,a relationship model between the characteristics of pedestrians and the number of pedestrians is established.(3)For different traffic situations,two optimizing timing schemes,timing control method and coordinated control method,are proposed.The effectiveness of the optimal timing control method is proved by simulation.