Research On Non-cooperative Target Capture Strategy Of Multi-manipulator System Based On Game Theory

More and more non-cooperative targets have the ability of autonomous movement with the rapid development of space technology,which leads it difficult to accomplish the capture task reliably depending on the traditional single-manipulator system.Benefit from the advantage in numbers,multi-manipulator system could surround the target to limit the escape probability.However,for these traditional target capturing strategies such as planning the end-effector of the manipulator to a predetermined position to capture the target,the game between target and multi-manipulator system causes poor adaptability and low success rate.Therefore,it has important theoretical significance and practical application value to research on non-cooperative target capture strategy of multi-manipulator system based on game theory.This research comes from the project of the Whole Design Department of Beijing Space Aircraft named "Development of XXX Control System for Capturing the XXX",and takes the multi-manipulator system as the research object and studies the target capture strategy based on game theory.The main work is listed as follows:Firstly,considering the multi-agent as pursuers,the game capture strategy of non-cooperative target with multi-agent is studies.A discretization method is used to establish the game scenario,and realize the target motion prediction based on the least square method,which lays the foundation for the research of the game capture strategy with multi-agent.By constructing the game model including players,strategy set and payoffs function,a real-time strategy selection algorithm is proposed to solve multi-agent capture strategy based on global optimum.Then an improved model is established by considering the limitation of the players'motion performance,improves the efficiency of game solving.Secondly,this paper extends the pursuers to the multi-manipulator system,and studies the game capture strategy of multi-manipulator system with fixed base.Aiming at the multi-manipulator system,the kinematics model and dynamics model of the multi-manipulator system are established,and a constraint set including workspace constraint,joint performance constraint and obstacle avoidance constraint are also established.And a multi-manipulator system game model is constructed to solve the game strategy.Furthermore,considering the joint torque optimization,the PSO algorithm is used to optimize the multi-manipulator single-step game strategy,and the execution strategy of the manipulators joints is further optimized.Thirdly,considering the complexity of the task,the game capture strategy of multi-manipulator system with controlled base motion is studied.Aiming at this problem,the relationship between base and target motion description is established,which lays a foundation for subsequent research.At the same time,a motion stage splitting method is proposed,which divides the capture process into the pursuit phase and the catch phase.In the pursuit phase,a one-to-one differential game model of the base and the target is proposed,and the decomposition method is used to solve the optimal pursuit strategy.And in the catch phase,a method of converting the game capture problem of base motion in inertial frame to the game capture problem of fixed base motion is propose,which reduces the complexity of the problem.Finally,the experiments of the non-cooperative target capture strategy based on game theory are researched.Reconstructing the ground experimental platform of space manipulators,experiments to validate the non-cooperative target capture strategy of multi-imanipulator system based on game theory is designed.The effectiveness and practicability of the proposed capture strategy is verified by analyzing experimental data.