Adaptive Cooperative Control Method Of Multiple Trains Based On Train To Train Communication

As a safe,reliable,efficient and convenient,energy-saving and large-capacity transportation mode,urban rail transit plays a key role in relieving the pressure of urban traffic.Its advanced automatic train operation control method is an effective means to ensure the safe and efficient operation of the trains.The automatic control process of urban rail transit train operation is inevitably affected by factors such as operational resistance,line conditions,and external disturbances,resulting in uncertainty.This requires significant self-learning and adaptive capabilities for its core control algorithm.In addition,the operation of urban rail trains is mainly based on the moving block mode and adopts the automatic driving and decentralized adjustment mode,which is difficult to upgrade additionally the operation efficiency of urban rail transit while guaranteeing the safety of train operation under the existing state of the short tracking interval,high driving density and dynamic coupling operation.As the optimal choice to improve the overall train operation efficiency and overall system security,multiple trains cooperative control based on train to train communication has become a research hotspot both domestically and internationally in recent years.This dissertation focuses on the abovementioned problems of the automatic train operation control and multiple trains cooperative control based on train to train communication,and mainly includes the following research contents and conclusions.(1)Considering comprehensively the influence of basic resistance,additional resistance,unknown external interference and internal factors on the train operation status based on train to train communication,based on the mechanism of motion mechanics,a reconstruction method for the uncertainty of train operation process based on RBFNN is established to compensate for the lack of self-learning ability of train dynamics models in changing operating environments.The safety and control accuracy requirements of multiple trains cooperative operation and station fixed point parking are analyzed in depth,and based on the theory of output constrained control and backstepping design method a train position output constrained control method that strictly limits the upper and lower bounds of the train operation position output values is proposed.And considering the uncertainty of train operation resistance parameters,an adaptive control method is designed to identify the resistance parameters,ensuring the good adaptability of the train operation control system.The results show that the train can track the expected train operation curve with a predetermined position tracking error of ± 0.21 m,fundamentally reducing the probability of multiple trains cooperative control position error “delay”.The overall speed error is between-0.24m/s ～ 0.26m/s,achieving precise tracking control of the target speed distance curve.(2)Considering the limited minimum safety interval of multiple trains cooperative operation based on train to train communication,an adaptive cooperative control of multiple trains based on RBFNN with constrained position output is proposed.Aiming at the cooperative control problem of state constraints and communication cooperation modes in the operation process of multiple trains,the impact of state constraints on the adjustment process of train cooperative operation is analyzed.An RBFNN nonlinear dynamic model for multiple trains is constructed,and a RBFNN position output constrained adaptive cooperative control method is designed in the predecessor-following train to train communication mode to adjust the speed and relative position of each train in real-time,ensure that the tracking interval between the tracking train and the preceding train is stable at the minimum safe position,and achieve the stability of multiple train cooperative operation queues.Further,the proposed cooperative adaptive control method is extended to the multiple trains cooperative control process in the lead-adjacency communication mode,and the multiple trains distributed cooperative control strategy is designed to ensure the stability of multiple trains cooperative operation.The results show that the leading train can track the desired train position curve with a pre-defined accuracy of ±0.21 m,and the safe distance between the rear train and front train is always 500±0.21 m,which can realize the stable and safe operation of multiple trains.(3)Considering the dynamic adjustment control process of multiple trains cooperative operation based on train to train communication,a based multi-agent adaptive cooperative control method with constrained output of train group position is proposed.The train group in the multiple trains cooperative control system based on train to train communication is regarded as a multi-agent system.According to the characteristics of constrained train operation position output,time-varying operation environment and uncertain model parameters,based on the realtime operation information of adjacent trains,the distributed position output constrained adaptive cooperative control method is designed to ensure that each train can track the expected operation track at the same time while maintaining the distance between any two adjacent trains within the designed safety range,achieving the safety of multiple trains cooperative operation.The results show that each train in train group position output constrained adaptive cooperative operation queue based multi-agent can track the ideal position curve with a predetermined position accuracy of ± 0.2m,and the speed of each train can reach consistency,thereby maintaining the train spacing at the set expected distance and ensuring the safe operation of multiple trains.(4)Considering reducing the computational complexity and control algorithm structural complexity of the train group position output constrained adaptive cooperative control system based multi-agent,a train group position output constrained adaptive cooperative control method with single value learning ability is proposed,which improves the structural complexity and engineering practicality of the algorithm while ensuring the performance of train cooperative control.The results show that the single value learning adaptive cooperative control with constrained multi-agent position output can effectively and adaptively estimate uncertain operating environments and changes in train parameters,and adjust its own operating status based on the operating information of adjacent trains.Under the premise of effectively ensuring the safety and stability of multiple trains cooperative operation,the algorithm’s structural complexity and engineering practicality are improved.This also provides reference for the research and application of multiple trains cooperative control systems under the new generation of train to train communication platform.