Research On Vehicle Information Detection System Using Geomagnetic Sensor

The rapid development of contemporary society and economy promotes the development of communication technology,information technology and vehicle industry.At present,vehicles have become an important transportation in modern society.With the rapid growth of its number,vehicles not only brings people comfort and convenience,but also makes traffic safety,urban congestion,environmental pollution and other problems increasingly serious.Intelligent Transportation System(ITS)is considered to be an effective means to improve urban traffic efficiency and driving safety.Intelligent transportation industry is a global innovation hotspot and an important commanding height of industrial development.ITS has a high requirement for the data collection of vehicle information,which requires collect huge amounts of vehicle flow and speed information on the road.However,the existing detection device of vehicle flow and speed is realized through video and radar,which only monitors the relatively important road sections,and cannot realize the large-scale data collection.In order to solve these problems,with studying on geomagnetic sensor,a low cost,wide coverage vehicle information detection system based on the geomagnetic sensor is proposed.This system can capture the magnetic field variation characteristics when the vehicle passes by at a lower cost,realize reliable vehicle information detection,and provide key data for the intelligent transportation system.The specific research content mainly includes the following two aspects:On the one hand,through the detailed study of the geomagnetic disturbance information generated by moving vehicles,a state machine vehicle detection algorithm with dynamic baseline update is proposed according to the information characteristics.Dynamic baseline update can adapt to the complex and changeable urban road traffic environment,and the state machine vehicle detection algorithm can judge the arrival and departure of vehicles quickly and efficiently.The Gaussian Mixed Model(GMM)is proposed based on the collected geomagnetic disturbances data during vehicle passing,and the ExpectMaximization Algorithm(EM)is used to separate vehicle data from background noise to obtain the best detection threshold.Compared with the existing works,the vehicle detection system studied in this thesis has the dynamic updated baseline,the appropriate threshold selection scheme and a more efficient state machine algorithm.On the other hand,starting from the actual deployment and application scenarios of vehicle detection,the problems and the information that can be used existing in the cooperation of multi-sensor are studied in detail,and the speed estimation algorithm of multi-sensor data association is proposed.Through this algorithm,the sensors with missed detection and redundancy detection can be found.For the sensors with missed detection,interpolation method is used to complete the data,and for the sensor with redundancy detection,the corresponding data is deleted.Compared with single geomagnetic sensor for speed measurement,the accuracy of our proposal is higher,which solves the problem of error data generated by missed and redundancy detection when a single geomagnetic sensor is used for vehicle detection.Finally,the stability and the accuracy of vehicle detection,the robustness of data association algorithm and the accuracy of vehicle speed estimation are evaluated through simulation and field testing.The results show that the vehicle detection system can work stably and deal with the changeable and complex urban road traffic environment.The data association algorithm can accurately correlate the vehicle data and find out the sensors with missed and redundancy detection.The vehicle speed estimation algorithm can accurately estimate the vehicle speed.The experimental results show that the accuracy of vehicle detection is 98.76%,and the accuracy of vehicle speed estimation is 98.21%.