Applications In Automatic Train Operation System With Adaptive Iterative Learning Control

When iterative learning control method deals with the tracking problem of the repetitive operating system,theoretically,it can eliminate the repettive error of certain system completely and realize the perfect tracking to the desired profile.However,the non-strict repetitive error is objective existing ineviatably in actual systems,including time-varying exterior disturbance,initial state shift,input constraints and state constraints.In fact,traditional iterative learning control method can not generate intime response to the non-repetitive disturbances.With the increasing of iterative times,the system's tracking error accumulates continuously.When the error accumulates to a certain degree,the instantaneous output will be too large and as a result,the system will lose stability.Eventually,the control peraformance of the systems will be influenced.Therefore,targetting the emerging problems of the iterative learning control study,this dissertation focuses on the automatic operating system of high-speed trains.The train follows the designed control laws in order to learn useful repetitive information during the process and improve the tracking accuracy to the desired operating profile.The main works and contributions of this dissertation are summarized as follows:1.This part discusses the adaptive iterative learning control problem of high-speed train operating process with time-varying exterior disturbances.This part separates the research of train resistance model into two parts: parametric model and non-parametric model.When dealing with the parametric model,recursive least square method in the iterative region is applied to identify the resistance coefficients;when dealing with non-parametric model,the proposed radial basis function neural network based on fuzzy inference mechnism is utilized to approximate the coefficients.Additionally,sliding mode control method is applied to compensate the approximate error and systematic diturbances.By designing parametric updating law in both iterative region and time region,the asymptotic convergence characteristic of tracking error along the iterative axis is proved based on the establishment of Lyapunov-like composite energy function.The error will convergent to a small neighborhood of zero when existing non-repetitive random disturbance.At last,the effectiveness of the proposed algorithm is verified by simulation analysis and experimatal test.2.This section focuses on the adaptive iterative learning control problem of high-speed train operating process with state constraints.First,the control problem of constraint system with initial consistency is studied.Then the initial state shift and exterior disturbances are taken into consideration.The saturation function is applied to restrain the upper bounder of system input and status.And the time-varying boundry layer is introduced to handle the raondom initial state.Then the adaptive control law and parameter updating law are designed based on Lyapunov function.Consequently,the asymptotic convergence along the iteration axis of the tracking error is proved through strict theoretical derivation.And the effectiveness of the proposed algorithm is verified by simulation and experiments.3.This part put emphasis on the adaptive iterative learning control problem of high-speed train operating process with initial non-uniform.At first,the desired error dynamic model is established according to the train dynamic model.In the meantime,by taking adequate consideration of time-varying equivalent mass into train operating process,the system tracking error is utilized to approximate the non-parametric uncertain model.After the preset time point,the error can convergent to an allowable adjustable region.Then,the robust iterative learning controller is designed.Consequently,the system can realize asymptotic tracking to the desired error profile along the iteration axis by strict mathematical justification.Also,the system control performance is improved under the interference of initial shift.