Study On Train Speed Fusion Method By Fuzzy Adaptive Federated Kalman Filter Algorithm

With the continuous improvement of running speed for the high-speed train,it has promoted the train control technology and performance indexes to develop in a higher direction.The train control system realizes the intelligent control of vehicle by monitoring the speed and position information of the train to ensure traffic safety and operational efficiency.The on-vehicle speed measurement module is the most basic and key sub-system of the train control system.And the speed information it provides is the decisive parameter to ensure the precise control of the train.Therefore,it is of great significance to improve the speed measurement accuracy of train control system and realize the autonomous control of the train.The current status and future development requirements of railway speed measurement system both at home and abroad are analyzed in this thesis.A combination of radar and GPS is introduced on the basis of the axle speed sensor to realize the speed measurement.After correcting the error of the axle speed sensor,its speed measurement value is combined with other speed information based on the fuzzy adaptive federated Kalman filter algorithm.The fuzzy comprehensive evaluation method is used to dynamically adjust the information distribution coefficients and realizes covariance shaping adaptive filtering,which can improve the accuracy,fault tolerance and reliability of the entire speed measurement system.The main research contents of this thesis are as follows:First of all,the principle,application status and error sources of axle speed sensor,Doppler radar and GPS are introduceed in this thesis,respectively.It shows the feasibility and effectiveness of the speed information fusion of the three sensors and establishes the basic framework of the speed measurement system for multi-sensor information fusion.Secondly,based on the widely application of axle speed sensor,its error is corrected to ensure the validity of the input data for fusion system.For the compensation of error caused by the wheel idling or sliding,a speed detection method is used and setting the detection threshold.The speed value of radar is compared with the speed value of axle speed sensor to judge the running status of the train wheel.If it is determined that the train is idling or sliding,the error is directly compensated by the radar speed value.For the correction of wheel diameter,based on the historical storage data,the intelligent method that FSA(Fast simulated annealing)optimized GPR(Gaussian process regression)is proposed to find the non-linear relationship between wheel diameter value and train running mileage,which can realize the automatic prediction and update the wheel diameter value.The experimental results show that the proposed algorithm can track the changes of wheel diameter values better than the other algorithms,and has higher prediction accuracy.Finally,the fusion algorithm of speed measurement system is studied.Because of the complex environment of the train running,the statistical characteristics of filter system will change,which leads to the filtering accuracy decrease or even divergence.From the perspective of adjusting the information distribution coefficients adaptively,this thesis uses Fuzzy comprehensive evaluation method to evaluate the filtering effect of each sub-filter,and then gives filter confidence and filter effect level of each sub-filter at the current moment,so that the information distribution coefficients are dynamically adjusted according to the filter confidence.When the sub-filter is in sub-optimal state,covariance shaping adaptive filtering is performed to improve the filtering accuracy,fault tolerance and reliability of the entire speed measurement system.Comparing with other fusion algorithms,the experimental results show that the velocity error and distance error of the fusion algorithm that this thesis proposed are minimal.And fusion effect of the system can be guaranteed under the changes of statistical characteristics.In addition,the speed measurement fusion scheme in this thesis can obtain better speed measurement accuracy than each single sensor,and can meet the requirements of the train control system.