Vehicle Detection Based On Cooperative Smart Road Studs

The traffic information is the basis of the intelligent transportation systems.Data analysis is related to the accuracy and rationality of traffic information.The SRS(Smart Road Stud)have advantages in terms of cost,power consumption,detection accuracy,and universality to the environment.The vehicle detection principle and the SRS system are introduced in this paper.This paper proposes a variety of methods for detecting vehicle flow including waveform extraction algorithm,vehicle position recognition algorithm based on geomagnetic theory and machine learning and cooperative SRS detection method.Using the above method,SRS can detect vehicles in parallel,cross-line vehicles and multi-lane traffic volume.Firstly,magnetic distribution around the vehicle is analyzed based geomagnetic principle,and it can be obtained the simulated waveform from the magnetic distribution mode of vehicle.The geomagnetic theory and the simulation provide a theoretical basis for traffic volume detection;In addition,the structure of SRS system and the way of data collection and transmission are introduced;Secondly,based on the principle of geomagnetic detection,a vehicle detection algorithm is proposed.The AMR triaxial geomagnetic sensor is used to collect the geomagnetic data of the vehicle.The geomagnetic detection principle is used to obtain the vehicle position discriminant feature.And the waveform extraction algorithm and the vehicle position discrimination algorithm are designed to complete a single SRS detection algorithm for the two-lane road;Thirdly,the construction and pruning principle of DecisionTree are introduced.After extracting the effective waveform features,the DecisionTree is constructed to determine the position of the vehicle.In addition,the unsupervised SOM clustering method is also introduced to determine the position of the vehicle.Both of which can accurately complete vehicle location identification and traffic volume detection;Finally,the detection accuracy of multi-lane vehicles is improved by SRS cooperation.The accuracy of the single SRS is limited by many factors.For example,the single SRS is impossible to accurately detect traffic volume when multiple vehicles are parallel.But multiple SRS cooperation can improve the detection accuracy and detect the traffic volume of multi-lane.In addition,cooperative SRS can distinguish the vehicles which cross lane line,multi-vehicle parallelism.So the Cooperative SRS Algorithm can adapt to the congested traffic environment efficiently.The vehicle traffic detection method proposed in this paper makes up for the defects of traditional traffic detection method,can accurately detect the multi-lane traffic flow,and can provide reliable traffic flow information for the intelligent transportation system.