Research On Traffic State Identification And Anti-collision Warning Based On DAS Technology

Intelligent transportation system(ITS)is an advanced transportation system that integrates advanced information technology,sensor technology,electronic control technology,system engineering and artificial intelligence technology into traffic management and ensure safety,improve efficiency,improve environment and save energy.The realization of intelligent control and dynamic safety warning of road traffic status is significant in intelligent traffic system,and the accurate detection and identification of vehicle traffic state is an important basis to achieve this goal.Therefore,the research on road traffic state identification and vehicle anti-collision warning has become an important topic in the research field of road intelligent traffic system.Distributed fiber optic acoustic sensing(DAS)that uses optical fiber as the sensing transmission medium is considered as the most advanced vibration acoustic sensing technology.DAS technology has obvious advantages such as simple structure,convenient use,wide measurement range,high sensitivity,large dynamic range,and antielectromagnetic interference.How to apply DAS technology to the reality of vehicle traffic state identification,and how to propose the corresponding model and algorithm of traffic state vibration signal processing to improve the accuracy and safety of traffic state data detection are seemed as an important research direction of road intelligent transportation system.Based on the distributed optical fiber acoustic sensing technology,this dissertation focuses on the problem of vehicle traffic state identification and the anti-collision warning strategy.This dissertation systematically explores how to solve the specific problems of vehicle traffic state identification,vehicle classification and vehicle anti-collision warning in traffic state identification based on DAS technology.Based on the traffic state theory and the complex signal analysis theory,this dissertation synthetically uses the methods of pattern recognition,parameter estimation and signal processing,and combines the system analysis with the case experiment simulation,the qualitative analysis with the quantitative analysis.Based on the requirement specificity and practical practice of the these problems,this dissertation proposes a vehicle traffic state identification model based on DAS technology,a vehicle type characteristic parameter identification model based on DAS technology,and a vehicle anti-collision warning strategy based on DAS technology.The main research contents and results of this article are as follows:(1)Research on traffic state identification model based on DAS technology.Different from the traditional vehicle traffic state identification model,this model based on DAS technology is to build a detection system and uses common optical fiber cable to detect the traffic vibration signal,and also it can receive fully distributed vibration information in the whole fiber link coverage area to detection and locate the vibration signal at any point in the coverage area.This model uses DAS technology to obtain traffic vibration data,and improves the wavelet threshold algorithm and the double threshold algorithm based on the characteristics of the vibration data.The former realizes the preprocessing of the vibration data,and the latter can process and analyze the vehicle traffic state and speed of the vibration signal.The experiment shows that in the vehicle counting test,the counting error of single vehicle passing through the detection area is small,when multiple vehicles pass through the detection area continuously,the counting error is large;In the speed estimation,the calculation result has a good accuracy,and the error is controlled within 5%.(2)Research on vehicle type feature parameter identification model based on DAS technology.According to the characteristics of DAS system that has low maintenance cost and can realize larger-scale vehicle classification data acquisition,this model uses sensing fiber to collect traffic vibration signals in the form of distributed sensors,and then extracts some features from the signals through the corresponding signal processing algorithm to identify vehicle categories.Based on the in-depth analysis of the vehicle classification problem,this model proposed the vehicle classification standard for this experiment scenario.In addition,in the signal processing step,a Support Vector Machine algorithm(SVM)based on genetic algorithm is proposed,and several SVM algorithms are compared in the form of algorithm comparison experiments.The results show that the improved SVM algorithm can reduce the classification error.In this scheme,a real scene is designed to test the proposed vehicle classification method.The results show that the accuracy of the DAS based classification detector is more than 70% in vehicle classification.(3)Research on vehicle anti-collision warning strategy based on DAS technology.On the basis of the research on the model of vehicle traffic state identification and vehicle type classification characteristic parameter identification,the vehicle anti-collision warning strategy is further studied.This section proposes a new vehicle anti-collision warning strategy,which is different from the traditional warning strategy.In this proposed strategy,the solution is a full-segment,full-scale vehicle anti-collision warning strategy based on the DAS system to detect road sections.Firstly,the vehicle safety distance model which is suitable for the vehicle collision warning strategy is established,and then a vehicle collision risk assessment method is proposed,which takes into account vehicle movement,vehicle position,driver behavior,road information and vehicle type.Finally,the simulation experiment is carried out to verify the accuracy of the proposed collision risk assessment method.The results show that the strategy can detect the collision risk effectively,and the give accurate collision warning in time.(4)Based on the actual research projects,the dissertation makes an empirical analysis of the proposed model and algorithm.The research results have important value for the establishment and improvement of the theoretical system of vehicle traffic state identification based on DAS technology,it provides a scientific basis for the actual planning and construction of road intelligent transportation system based on DAS technology.