Research On Vehicle Detection Scheme Based On Microwave Sensors Collaboration

With the booming development of economy and automobile industry,the vehicle ownership is increasing dramatically,and fully autonomous driving is expected to be realized in the future.However,a series of problems such as traffic congestion,resource consumption and environmental pollution are becoming more and more serious.Moreover,the reliability of autonomous vehicles in complex environments still needs major improvement.To solve the above problems,intelligent transportation systems(ITS)connects roads,vehicles and pedestrians closely,and promots the construction of intelligent transportation greatly.As the foundation and the key component of ITS,vehicle detection is able to provide indispensable traffic information for traffic surveillance and driving decision,which is essential for the construction and development of ITS.On the one hand,vehicle detection is often used to collect vehicle information at monitoring points.On the other hand,vehicle detection is also widely used to obtain real-time vehicle information of a specific section or the whole road.To provide real-time and reliable traffic information for traffic surveillance and driving decision,a cost-effective vehicle detection scheme based on Internet of Things(IoT)technology is proposed for urban road scenarios where the traffic flow,the speed and the type of vehicles change greatly.Different from the existing complex and expensive vehicle detection schemes,the proposed scheme has innovative advantages.The performance defects of low-cost sensors are overcome,and vehicle detection is realized with high accuracy.The scheme is suitable for large-scale application scenarios to realize regional perception with wide coverage.In view of the above analysis,the main work of this thesis is as follows:On the one hand,after investigating the weaknesses of current vehicle detection schemes,this thesis balances the cost and detection accuracy of vehicle detection at monitoring points.Low-cost vehicle detection nodes are deployed on the roadside to collect the data of passing vehicles on the roads.Then,to avoid processing microwave signals in frequency domain in the traditional methods,a dual-node collaborative vehicle detection algorithm based on state machine is proposed to realize vehicle counting,speed measurement and vehicle classification.After this,field experiments are conducted in different scenarios to evaluate the proposed algorithm,which can overcome the performance defects of low-cost microwave sensors and satisfy the needs of vehicle detection on urban roads with the unique advantages of low cost and high accuracy.On the other hand,to satisfy the needs of obtaining real-time vehicle information of a specific section or the whole road in large-scale IoT application scenarios,a large number of vehicle detection nodes are utilized to realize regional perception with wide coverage,and a multinode collaborative vehicle detection strategy is proposed.The best function match of the data is realized through least squares method,and the system state is predicted and updated in real time based on the extended Kalman filter.The proposed strategy overcomes the accuracy limitation of dual-node detection algorithm.Finally,according to the data collected in urban roads scenarios,simulation data are generated and simulation experiments are carried out.The results validate the high accuracy and stability of proposed strategy,which is suitable for large-scale IoT application scenarios and can provide real-time and reliable vehicle information for traffic surveillance and driving decision.