Optimal Control Of Automatic Train Operation System Based On Adaptive Dynamic Programming

The rapid economic development has brought about increasing traffic demands and serious traffic problems.The rail transit system provides a solution to it.The automatic train operation(ATO)system can improve the operating efficiency of the rail transit system under the precondition of ensuring the train traffic safety.At the same time,with the development of intelligent technology and computer technology which provides an effective solution to the multi-objective control of complex system,it becomes possible to find an efficient and reliable control method of automatic train operation system.The ATO system generates the speed-distance train operation target curve based on optimization algorithms to meet the demand of energy saving and punctuality according to the line data and train parameters.And then,by using efficient control methods,ATO ensures that the train operating in accordance with the target curve maximally,and achieves satisfactory control effect.In this paper,we take train traction energy consumption and train operation time as the main optimization goal.Based on the advantage of dynamic programming by which we can obtain the global optimal solution,we firstly divide train operating intervals and build energy consumption matrix and operating time matrix during the operation process of train,and then,the train driving strategy sequence is iterated offline,finally we get the speed-distance train operation target curve which satisfies the energy saving and punctuality.And the problem of adjusting the curve of the target curve when the lights change in the running process of the train is also studied.Considering the curse of dimensionality and time reversal,dynamic programming is not suitable for real-time control of complex systems,but adaptive dynamic programming is a solution to it.Based on the generated target curve and the adaptive dynamic programming method,the ATO control algorithm is studied.We achieve satisfactory speed control accuracy and parking accuracy while effectively reducing the oscillation of the system which may occur during the tracking of the target curve.At the end of this paper,the method adopted in this paper has been verified by simulation to prove its feasibility and effectiveness.