Research On Expressway Event Detection Based On Multi-sensor Data Fusion

Expressway incident detection is one of the important directions in the research of intelligent transportation systems and is also the main research content in the field of traffic information processing.China's expressway accident detection research has been connected with the international track since 1990.The accuracy of incident detection results has always been the focus of experts in road traffic research.At present,because the data fusion has not achieved satisfactory results in the traffic field,establishing a highway event detection method based on video and binary sensor data fusion can effectively solve the problem of poor accuracy of a single data source.This article explores the key technology research of high-speed smart transportation in Urumqi around the city based on the Internet+risk environment information service project of the Chongqing Science and Technology Commission project,and based on the theory and technology of environmental information collection,proposes an accurate traffic status parameter.The sensor design method based on binary detectors,at the same time,taking into account the impact of special meteorological conditions,design the threshold correction algorithm to improve the traditional event detection algorithm,and finally,using the VISSIM simulation tool to compare the detection results of the fusion method and single algorithm.The main contents of this paper are as follows:(1)Firstly,two aspects of literature investigations were carried out for traffic incident detection algorithms and multi-sensor data fusion methods.Grasping the advantages and disadvantages of all kinds of traffic detection technologies and aiming at many factors such as the difficulty of highway event detection and high cost,a method for acquiring traffic state parameters based on binary detectors was proposed,and it was proved through experimental simulation that within a certain range virtual serial port simulation values and theoretical values are basically the same.(2)Secondly,the threshold correction algorithm is designed to improve the traditional event detection algorithm.Through the discussion of the traditional algorithm theory,the problems of the traditional event detection algorithm under special weather conditions are found.The concept of rain and fog detection was introduced to include rain and fog information in decision making.Considering the factors of cost anddifficulty of deployment,a sparse layout scheme and information processing method for detecting the rain-flood detectors in the road section are proposed.The low flow algorithm corresponding to different traffic conditions and the threshold correction algorithm of the standard deviation algorithm are given.Data validation was performed on the algorithm.The results showed that the threshold correction results were basically consistent with the experimental simulation results.Appropriate threshold correction can reduce the impact of severe weather conditions on the detection results,and the accuracy of the detection algorithm will also be improved.(3)Finally,the voting fusion method is introduced into the field of event detection and a highway event detection method based on multi-sensor data fusion is proposed.The method effectively completed the automatic identification algorithm based on video detector module and an improved algorithm of low flow based on binary detector or standard deviation algorithm module in the fusion of policy makers.Simulation of the expressway through VISSIM traffic modeling and simulation software shows that the voting fusion method is superior to the single algorithm in event detection performance.It can not only make full use of the advantages of the single algorithm,but also improve the detection rate of events.