Train Speed Curve Tracking Study Based On Iterative Learning Control

The key of train control system is adjusting the speed of the train automatically by theintelligent algorithm, and the research is one of the main functions of automatic trainoperation (ATO). At present, the high-speed railway is artificially driving mode. Theexperience of the driver’s controlling the train is not the same,which has an impact on thecontrol effect. In the high-speed railway system, designing the train speed controller to trackthe target speed curve optimized has important significance. The main contents are below.Since the same technical principle between the urban rail transit system and ChineseTrain Control System-3(CTCS-3), the specific plan based on these equipments which isadding speed controller module on CTCS-3system is proposed to complete train autopilotand information interactive function.There are a lot of useful information contained in the running are ignored, and the trainwill be subject to multiple effects from themselves and the environment. It is difficult toestablish a precise mathematical model.The train speed controller is designed by usingiterative learning control (ILC) which only needs the historical input and output data of thesystem. And the disruption running can be turned into load disturbance.So the design canreduce the impact of interference on the system, and improve control accuracy of the system.Target speed of the train is only based on the position profile and the current speed of thetarget position.The actual speed of the train without full consideration of the target during theoperation is required to meet safety, punctuality, comfort, high precision and low powerconsumption requirements.Optimizing the curve by particle swarm optimization (PSO), thecontrol curve can be smooth and avoid repeated switch between acceleration and deceleration,reduce energy consumption, improve comfort,and provide the best control strategy for thetrain to run.In order to verify the reasonableness of the design, using VC++6.0to come truesimulation experiment, and designing the control systems and human-machine interface.Upon completion of the initialization data to control the train running by the optimized targetvelocity curve, the train can adjust the speed automatically, reducing the labor intensity ofdrivers, and then analyzing the performance of the train control mode, judging the merits ofthe control strategy, and providing a priori conditions for future technological developments.