| If the environment was not changed,traditional manipulator can adjust the parameters to achieve the required goals according to the content of the tasks.Due to the limitations of manipulator configurations,it is very difficult to complete many tasks in the premise of unchanged manipulator configurations,and designing different configurations for different tasks will increase the cost.Finding a manipulator that can meet the needs of different tasks and change its configuration to adapt to environmental changes is an urgent need for people,and the reconfigurable manipulator was born.Reconfigurable manipulator has great development prospects in many application fields such as industrial manufacturing,lunar exploration,medical industry,etc.In addition,it has many advantages such as high flexibility,low cost and strong adaptability,etc.,and it also has important research value on the enhancement of the precision of the reconfigurable manipulator system,optimization of the controller structure and reduction of energy consumption.Therefore,focusing on the research of reconfigurable manipulator has far-reaching theoretical significance.Firstly,this paper describes the background and significance of the topics selected,and reviews the research status and research content of reconfigurable manipulator at home and abroad,and summarizes the adaptive dynamic programming(ADP)methods,then elaborates the full text of the chapter arrangements and the main research content.Secondly,the Newton-Euler equation of rigid body is analyzed in the object coordinate system,and according to Newton-Euler iterative algorithm,the generalized speed and generalized acceleration of the connecting rod are used to carry out the positive iteration between the modules,and the generalized force is iteratively iterated between the modules.Finally,the dynamic model of the reconfigurable manipulator is derived.Thirdly,considering the modularity and reconfigurable characteristics,decentralized control method is more suitable for reconfigurable manipulator system,and the dynamics model of reconfigurable manipulator subsystem is deduced,and the reconfigurable manipulator decentralized tracking control method based on adaptive dynamic programming algorithm is proposed.By neural network theory,critic structure with a single network is used to evaluate the structure as an agent,and the solution of Hamilton-Jacobi-Bellman(HJB)equation is deduced by online strategy iterative algorithm,then get the optimal feedback control strategy and the decentralized optimal control strategy,and the stability analysis and simulation verification are carried out.Fourthly,the dynamic model of reconfigurable manipulator subsystem based on joint torque perception is established,and a learning-based decentralized optimal control method is proposed.Through the feedback control strategy,the decentralized optimal control problem of the system can be transformed into the optimal compensation problem of the model uncertainty to deal with the interconnected coupling terms.The HJB equation is solved by the strategy iteration method and the critic network is constructed to approximate the optimal feedback control strategy,and the effectiveness of the proposed control method is verified by simulation.Finally,the work carried out in the full text is summarized,and combined the problems and challenges in the study,and the next step of research work is forecasted. |
PDF Full Text Request