Study On The Automated Identification Of Traffic Abnormity State For Arterial Roadways

At present, with traffic congestion emerging frequently and accidents beingsevere increasingly which become the difficult challenge all over the world, severalproblems have been brought about, such as traffic delay increasing, traffic safetydecreasing and cost wasting hugely, which have been disturbing the living ofpeople more and more. To alleviate the negative effect of congestion and accidents,it is very necessary to provide the traffic managers and the travelers with theinformation on the traffic abnormity state newly and reliably. If the trafficabnormity which is on the point of emerging and has been emerged could beidentified accurately, adopting the effective traffic management measures mayreduce the severity extent and even avoid the traffic congestion happening.The thesis stems from the project key theories and simulation technologies ofdynamic traffic management and control in urban road network, funded byNational Natural Science of China under grant 50338030. From the perspective oftheory and practice, aiming at insufficiencies of the research on the automatictraffic abnormity identification in the urban road network at present, such as thelimited subjects of developed traffic abnormity identification algorithms and theundesirable performance of algorithms, the thesis experiments the advancedinformation processing technologies including Artificial Neural Network andStatistical analysis, respectively proposes the automatic recurrent congestion andnon-recurrent congestion identification algorithms using fixed detector and probevehicle data, and the automatic traffic accident detection algorithm in the light flowfor arterial roadways.The thesis comprises of 6 chapters and their contents are as follows:Chapter One: Introduction. First, present the purpose and meaning of research.Then, the main content and structure of the thesis is given.Chapter Two: the presentation of the research background. Firstly, summarizethe background of the automatic traffic abnormity identification for arterialroadways and introduce its history and status quo. Secondly, define the concept oftraffic abnormity state in detail. Finally, present the index and method evaluatingthe performance of traffic abnormity identification algorithms.Chapter three: the design of automatic traffic congestion identificationalgorithms using fixed detector data for arterial roadways. First, the thesis reviewsand summarizes the general traffic congestion identification algorithms. Based onthis introduction, the automatic non-recurrent congestion and recurrent congestionidentification algorithms using fixed detector data are proposed, and that theirdetailed design process is presented. After that, the thesis calibrates and tests thenew designed algorithms using simulation data and field data and shows thevalidity of algorithms by obtaining the performance index.Chapter four: the design of automatic traffic congestion identificationalgorithm using probe vehicle data for arterial roadways. After reviewing thegeneral identification algorithms using probe data, the detailed design process ofalgorithm based on the travel time of link and travel speed of link is available.Finally, simulation data are used to calibrate and test the innovative algorithm.Chapter five: the design of automatic traffic accident detection algorithm inthe light volume for arterial roadways. First, look back the general algorithms andpropose the innovative detection algorithm based on the extent of vehicleacceleration change. Then, calibrate and test the new algorithm using simulationdata and field data.Chapter Six: Conclusions and the future work. The main contents of the thesisare reviewed, and then the innovation and the existing problems are pointed out.Finally, the problems that need to further study in the future are proposed.