Research And Design Of Overloading Detection System Based On Multi-sensor Information Fusion

As a powerful equipment of governance vehicle overloading,non-stop overrun detection system has been used in highway in our country for almost ten years,and has achieved good results.However,there are still problems in accuracy and reliability of the measurement in recent years.For this reason,the paper designed a vehicle overloading automatic detection system based on multi-sensor information fusion.The system is capable of detecting the vehicles under the condition of normal driving state.The system includes dynamic weighing subsystem,license plate number recognition subsystem,vehicle separation subsystem,vehicle identification subsystem,speed and overrun detection subsystem,limit indication and alarm subsystem,computer management subsystem and so on.The vehicle identification subsystem,speed and overrun detection subsystem,computer management subsystem are developed by our research subject.This paper mainly introduces the multi-sensor information fusion technology and its application in overloading detection system.The Bayes parameter estimation used for selecting weighing data,Kalman filter used for weighing data optimization,Production rules used for vehicle identification,Correlation used in speed detection are introduced in detail.This paper carried out research in the following aspects and has made innovative developments:1.The multi-sensor information fusion method including the Bayes parameter estimation,Kalman filtering,Production rules,Correlation algorithms applied to highway overloading detection system.The Bayes parameter estimation used for selecting weighing data,Kalman filter used for weighing data optimization,Production rules used for vehicle identification,Correlation used in speed detection and so on.2.The fusion Bayes parameter estimation using for t he dynamic weighting data sub-period.This principle is divided into three parts:confidence distance theory,the best fusion number selection method,calculation method based on Bayesian estimation fusion.3.Kalman filter is optimized to the vehicle weight data which after bayes estimation of two-channel sensor fusion were superimposed to obtain reliable weighing data.4.The Production system data fusion used in axle type identification system.After studying the common rules and characteristics axle type in the highway,designed three level rulebase using Production rule form,to achieve the fusion of axial type identification.5.In the speed detection System which based on dual laser sensor and Correlation algorithm,the left and right sensors are installed in two planes which are perpendicular to the direction which the traffic travels,two planes separated by a short distance.When a car passes through under the frame,first through under the right sensor's scanning plane,then through under the left sensor's scanning plane.Cross-Correlation algorithm basic principle is the use of the basic principles of the calculated Correlation function method to calculate the transit time of the upper sensor and the lower sensor signals,then calculate the velocity by transit time.6.The speed,length width height detection and non-stop fast detection system for fusion,to achieve the goals of comprehensive governance overloading and scientific governance overloading of the speed and dimensions,which makes the whole system function more perfect,widely governance irregularities overrun phenomenon.This design is installed and used in highway junctions of Tai-Chang,Taiyuan,Datong and so on in Shanxi,and has made remarkable achievements.Practice has proved,this design improves the weighing precision,reliability and work efficiency.Effectively eased the traffic jams of the entrance of the highway,bring great convenience for governance the vehicle overloading,to achieve the purpose of this research projects.