Research On Rapid Exponential Stability And Optimal Control For Reconfigurable Manipulator

In recent years,with the development of science and technology and the progress of technology,robot technology,and its application research has developed in full swing worldwide.In the US National Strategic Plan for Advanced Manufacturing,German Industry 4.0,Made in China 2025,Japan's New Robot Strategy,and Korean Robot Future Strategy 2022,robots have received unprecedented attention as the core technology.The reconfigurable manipulator is a special kind of robot system.The difference from the traditional manipulator is that it is composed of several modules with independent control and sensing capabilities.According to different task requirements and field environments,the internal modules can be local relative movement occurs between them,changing the configuration and changing the connection orientation,thereby changing the overall structure and function of the system.Relying on the outstanding characteristics of strong robustness,high flexibility,flexibility,and cost saving,when applied to environmental operation scenarios such as flexible multi-process processing,intra-cavity micro-surgery,rescue,and emergency outbreaks,the reconfigurable manipulator is more incomparable advantages.Many scholars have carried out in-depth research on issues such as module design and configuration optimization,automatic kinematic modeling and trajectory planning,trajectory tracking controller design,and fault diagnosis,and have obtained a series of rich theoretical research and practical results.However,the process of analysis and synthesis of the reconfigurable manipulator system behind the development of the prototype.The analysis and design using traditional robot theory can not meet the needs of technological development.This paper takes the reconfigurable manipulator as the research object,mainly focusing on several key issues of rapid exponential stability,optimal path planning,robust stability,trajectory tracking control and fault-tolerant control of the reconfigurable manipulator.Reconstructed manipulator system represents a class of nonlinear exponential stability analysis theory for nonlinear dynamic systems,forming a unified framework for the rapid exponential stability analysis of such systems,improving the efficiency of reconfigurable manipulators and realizing reconfigurable manipulator low energy consumption,high efficiency and stable working condition.The specific research content includes the following aspects:The problem of rapid exponential stability analysis of reconfigurable manipulator system is studied.The stability problem is the prerequisite for completing all tasks,and accidental or continuous interference is unavoidable in various application scenarios.For nonlinear,strong coupling and other characteristics of the reconfigurable manipulator system dynamics model,to start with rapid exponential stability analysis of the system,establish a theoretical framework for rapid exponential stability analysis;Propose a rapid exponential stability criterion for nonlinear control systems based on nonlinear system fast exponential stability,design an optimal feedback controller;Performance conditions and HamiltonJacobi-Belman equations,a rapid exponentially stable criterion for reconfigurable manipulators is proposed,and an optimal feedback controller is further designed.The robustness and real-time performance of the reconfigurable manipulator control system are improved,and the analysis method is extended to nonlinear hybrid systems and nonlinear stochastic systems.The optimal path planning method of reconfigurable manipulator system is studied.Consider the complex environment and real-time working conditions encountered during the actual operation of the reconfigurable manipulator system,analyze the kinematic constraints and obstacle avoidance strategies of its modular joints,and transform the path planning problem into a time-varying nonlinear optimization problem.As the shortest criterion for energy consumption evaluation,an optimal path planning method based on zeroing neural network is proposed.By constructing a type of zeroing neural network,designing an objective function that satisfies the task reachability,finding the shortest safe path of the reconfigurable manipulator,and theoretically strictly proves that the proposed zeroing neural network method has good stability.Finally,the problem of path optimization of reconfigurable manipulators under general non-structural environment constraints is studied,which ensures the minimum energy consumption while improving the task execution efficiency of the reconfigurable manipulator system and the adaptability to the working environment.The trajectory tracking control method of the reconfigurable manipulator is studied.Aiming at the problem of trajectory tracking control in the joint space of a reconfigurable modular manipulator with uncertainty and external interference,a PD-type accelerated iterative learning control method based on anti-interference estimation is proposed.The exponential variable gain is used to accelerate the learning control law and improve the convergence rate of iterative learning.Based on Lyapunov stability theory,the asymptotic stability of the reconfigurable manipulator system is analyzed.Based on the robust stability analysis of the asymptotic convergence of the nonlinear dynamic system,the criterion of the asymptotic stability of the periodic motion of the reconfigurable manipulator is studied;the joint torque of the reconfigurable manipulator system is obtained by using the joint torque sensor information,combined with the terminal sliding mode design controller idea,design an adaptive terminal sliding mode decentralized controller to solve the problem of independent joints of the unit joint of the reconfigurable modular manipulator system without using other joint information control.Based on the terminal sliding mode decentralized control,the RBF neural network is used to estimate and compensate the model uncertainty caused by the reconfigurable manipulator friction and interconnection terms,and further strengthen the control of the reconfigurable manipulator system.Real-time and strong robustness.Finally,a 2-DOF reconfigurable manipulator experimental platform with decentralized control architecture was built to verify the feasibility and effectiveness of the PD-type accelerated iterative learning control method and decentralized adaptive trajectory tracking control method.The fault-tolerant control method of reconfigurable manipulator system is studied.Aiming at the independent joint fault subsystem of the reconfigurable modular manipulator,the system actuator failure model is established,the adaptive neural network method is used to estimate the uncertainty of the model,and the sliding mode observer is designed to track the fault information,and then the design is decentralized.The control law reconstructs the controller.Formulate an active fault-tolerant control strategy.When the system is fault-free,adopt a nominal system decentralized control law;when the system detects a fault,immediately switch to the fault-tolerant control law based on sliding mode observer to reduce the lack of a priori knowledge in control hidden instability caused by reconfigurable manipulator system.The stability of the control system is analyzed using Lyapunov equation.Finally,the 2-DOF reconfigurable manipulator experimental platform verifies the feasibility and effectiveness of the decentralized active fault-tolerant control method based on sliding mode observer.Finally,the full text work is summarized,and the follow-up research work is prospected.