Distributed Control For High-Speed Train Based On Multi-agent Consensus Technology

Nowadays,high-speed railway transportation plays a more and more important role in social economy.Compared with traditional trains,high-speed train is more rapid,economic and environmental.However,the requirement of control system is higher.Traditional centralized control method are widely used,but the control effects still need to be improved.Therefore,it is of great significance to build a multi-particle model of high speed train in order to describe its dynamic characteristics more accurately,and to study more effective high-speed trains' control strategy to achieve accurate tracking.In this thesis,based on the deeply research on the present situation of high speed trains,the modeling and control methods of high speed trains are introduced in detail,and then control strategies for a multi-particle model are proposed.This thesis respectively on the design of distributed controllers for the system model of a single train and multiple trains based on the multi-agent consensus technique,the main research contents and results are as follows:The distributed cruise control problem of a single high-speed train's movements is investigated.Firstly,the problem of a single train with multiple cars is interested,and the aim is focused on the design of a distributed controller,with which the train achieves the safe moving,i.e.,all the cars track a desired speed and any neighboring cars keep a safe distance from each other.By considering,among the cars,the physical coupling structure between each neighboring cars constitutes a physical connected interaction graph,and by posing the speed information exchange mode that the communication topological graph has a directed spanning tree,a new approach based on the complete consensus technique is proposed to solve the problem.Secondly,the problem of the multiple trains with multiple cars is considered,and the attention is the design concentrating on a distributed control scheme such that,guaranteeing the safe moving of each single train,the multiple trains can track their respective speeds while simultaneously avoiding collisions.A new approach based on the cluster consensus technique is developed to solve the problem.Compared to the existing centralized control scheme or the decentralized control scheme,where the control design of each car need the measurement information of all the cars,the designed distributed control laws have the merit that the control design of each car only uses the measurements of itself and its neighboring cars;besides,the problems are solved in terms of the purely graphic topological conditions,which are easy to be verified.The final simulation results demonstrate the effectiveness of the design algorithm.A multi-agent system model of high speed trains is established and we design two consensus protocols according to its properties and consistency analysis.The stability of the system is proved by theoretical analysis,and the effectiveness of the proposed algorithm is verified by simulation results.This thesis will provide some theoretical basis for the further research of distributed control algorithm for high speed trains.Finally,the thesis makes a summary of the research and points out some open problems to be explored in future.