Researh Of Freeway Traffic Incident Detection Based On Adaboost And Pso-RBF

Highway, as the ties of leading modern culture exchanges and promoting rapid economic development, has been developed fast and successfully, and brought great convenience to our people in China. At the same time, its safety, open, efficient and other issues also attracted our widespread attention. Due to the complexity and particularity of modern highway, once there something occurs in any form of traffic incident, its capacity will be severely hampered, and even it will make a big threat to people’s lives and property safety. If we don’t take any measures on time, the result would be worse. Therefore, how to detect the traffic incident quickly and accurately, and take reasonable and effective measures to minimize the loss is always been studied in our society. Traffic incident detection algorithm, as the core of traffic incident detection, its overall performance is directly related to the stability of the entire detection system. Therefore, the study of traffic incident detection algorithm is very significant.The establishment of traffic incident detection model is based on the up-downstream traffic flow parameters changes where the traffic incident happened. This paper also follows the line of thinking, based on the sudden change principles of traffic flow parameters, designed RBF traffic incident detection model. The confirmation of RBF network parameters is relied on human experience and other factors excessively, here we choose PSO algorithm to improve RBF network parameters optimization; As the basic PSO algorithm has the problem of lower precision when optimizing, so this paper takes different ways to improve the standard PSO algorithm, and establishes IPSO-RBF-based traffic incident detection model with good detection results. Finally, according to the principle of "the more the difference of weak classifiers, the better the integration result", we use different combinations of traffic flow parameters to constitute IPSO-RBF weak classifiers in different network structures, and put in use of the AdaBoost method to integrate improved IPSO_RBF model, and build model to detect traffic incident.This paper analyzes different ways of traffic data collection, and ultimately designed and implemented based on video image to collect traffic data approach. Take I-880traffic data as the basis for simulation using MATLAB R2013a platform. By analyzing, the simulation results show that the method based on AdaBoost traffic incident detection integrated IPSO-RBF model has fast convergence and high precision classification and achieved good results in highway traffic incident detection.