Research On Vehicle Information Detection System Based On Geomagnetic Sensor

The vehicle information detection technology is the core technology to realize automatic control of traffic signal lights and non-stop charging of highway,which can provide information support for the application of intelligent transportation systems,Therefore,the further study of vehicle information detection technology has great significance to the development of intelligent transportation systems.In this paper,a vehicle information detection system based on geomagnetic sensor is designed.A state machine vehicle detection algorithm based on dynamic threshold and a vehicle type classification algorithm based on data fusion are proposed to complete vehicle information recognition.Experimental results show that the system can effectively improve the accuracy of vehicle detection and vehicle type detection.The main work of this paper as following:(1)This paper studies the geomagnetic disturbance information of moving vehicle collected by geomagnetic sensor and analyzies characteristics of disturbance information,we proposed the state machine detection algorithm based on dynamic threshold.The algorithm uses the weighting function to update the threshold in real time,which can eliminate the influence of baseline drift on vehicle detection accuracy.The state machine can effectively determine the time point when the vehicle enters and exits the sensor detection range,accurately segment the vehicle geomagnetic disturbance data,and improve the recognition accuracy of vehicle type,vehicle length and vehicle speed.(2)In order to solve the problem that has less vehicle information and low accuracy detected by the single geomagnetic sensor,multi-sensor node fusion detection is used to realize the fusion processing of multi-node vehicle geomagnetic disturbance perception.In this paper,a multi-node disturbing data feature alignment algorithm based on DTW(Dynamic Time Warping)is proposed,Adaptive weighted fusion of multi-node perceived data after regularization by correlation function,to complete multi-node vehicle geomagnetic disturbance data fusion at the data level.(3)According to the collected vehicle data,the detection vehicles are divided into small cars,SUVs,medium-sized trucks and large passenger cars.Feature vectors for vehicle classification are extracted from the merged vehicle geomagnetic disturbance data.In order to solve the problem that BP(Back Propagation)neural network has slow convergence rate and is easy to fall into local optimum,the GA-BP(Genetic Algorithm)neural network algorithm is proposed for vehicle classification.(4)This paper build the experimental platform to verify the vehicle detection and vehicle classification algorithm.Experimental results show that the accuracy of vehicle detection algorithm proposed in this paper is better than 98%,and the classification accuracy of vehicles is more than 84%,and data fusion can significantly improve the classification accuracy of vehicles.