Research On Tracking Control Of Virtual Rail Train Based On RBF Neural Network

With the continuous advancement of urbanization,many urban problems have become more and more serious.Virtual track train has entered the public view because of its advantages of environmental protection and large passenger capacity.As a new intelligent transportation tool for the future,virtual track train needs to have independent tracking function,which depends on its key tracking control technology.Aiming at the tracking control of virtual track train,this thesis mainly completes the following research work:Firstly,the dynamic model of virtual track train is established.Based on the investigation of the structural characteristics of articulated road vehicles and virtual track trains at home and abroad,the overall structure of virtual track train with three vehicles is determined,which is all-wheel independent drive and steering.The whole train system is divided into powertrain system,steering system,suspension system,braking system and tire module.The relevant parameters of body size and each part of the system and module are set in Truck Sim software.Then,the tracking control method of virtual track train is studied.Based on the investigation of relevant vehicle tracking control technologies at home and abroad,combined with preview control,a tracking control method based on RBF neural network PID theory is proposed,which integrates the advantages of RBF neural network’s powerful self-learning and adaptability and PID’s simplicity and reliability.So as to improve the tracking effect of the train,RBF neural network is used to adjust PID controller parameters online,so that the controller can flexibly adapt to different tracking conditions.Aiming at the problem of lateral movement of the train,a wheel angle control method based on Ackerman steering principle is designed.On the one hand,the wheel angle is reasonably distributed to ensure the tracking accuracy and steering stability of the train.On the other hand,coordinating the motion relationship between the vehicles and reducing the interference of the vehicles can avoid snake motion.Aiming at the longitudinal movement of the train,a wheel speed control method based on the same instantaneous center principle is designed.On the one hand,the wheel speed is reasonably distributed and the steering is assisted by differential drive to ensure the stability of the train.On the other hand,the running speed of each vehicle is reasonably configured to reduce the squeezing effect between vehicles and avoid the folding instability of the train caused by the running problem.Based on the first vehicle of virtual track train,the tracking control method are preliminarily verified.Afterward,the tracking control method of virtual track train is verified by simulation.Based on Truck Sim and MATLAB / Simulink,a joint simulation platform of virtual track train dynamics is built.The pass ability of minimum curve is calculated according to the overall dimension of train body and road width.And a reasonable target trajectory is designed.In order to verify the control effect of the control method,research on tracking simulation are carried out under different road conditions.In order to verify the anti-interference ability of the tracking control method,research on anti-interference simulation are carried out in different external environments.The simulation results show that the adaptive tracking control method based on RBF neural network PID proposed in this thesis can coordinate the motion relationship between vehicles and reduce the tracking deviation of the train and ensure the stability of the train in the tracking process.At the same time,it also has certain robustness and can adapt to the changes of the external environment.Finally,the feasibility of the tracking control method is verified based on the principle vehicle of the first vehicle of the virtual track train.The longitude and latitude coordinates of the target trajectory are collected by GPS equipment.And the longitude and latitude coordinates are converted into XY geodetic coordinates required by the control algorithm through Gaussian Kruger projection,which is used as the target trajectory of tracking test.Based on Lab VIEW software,an upper computer tracking control system is established,which includes four modules: GPS/INS equipment data receiving and conversion,control program loading,motor communication and experimental data acquisition.On this basis,under the speed conditions of 5 km/h and 7 km/h,the tracking test of the principle vehicle of first vehicle of the virtual track train with four-wheel drive and four-wheel steering is carried out,which proves the feasibility of the tracking control method proposed in this thesis from the practical application level.