Research On Optimal Fault Tolerant Control Method Of Multi-fault Concurrent Modular Manipulator Based On Adaptive Dynamic Programming

With the introduction of Industrial 4.0 and intelligent manufacturing,the research climax of modular manipulator has been set off.The modular manipulator consists of the joints and connecting rods of the unified standard physical and electrical interface,which can be independently altered to complete the needs of different tasks.Therefore,modular manipulator is widely used in the task of non-single,complex working environment,its sensors,actuators and other electronic components inevitably failure,and can not be human intervention,the next will result in significant property damage,even casualties and other major accidents.Therefore,it is of great theoretical and practical significance for the modular manipulator system to design fault-tolerant control system to ensure that the modular manipulator system can maintain certain control performance after the failure.In addition,for the module manipulator control system,combined with the optimal control theory,in the spirit of improving control accuracy and optimizing energy consumption research is also of great and far-reaching significance.From the current data,combined with the optimal control idea,the study of fault tolerance control methods for multi-fault concurrent modular manipulator is rare and not in-depth enough.In this paper,an optimal fault tolerance control method based on adaptive dynamic planning is proposed for the multi-fault concurrency of modular manipulator,and the optimal fault tolerance control method based on ADP's modular manipulator under the concurrent of sensor and actuator faultised is mainly studied.The main contents of the text are as follows:(1)Points out the research background and significance of the subject,and summarizes the theoretical knowledge of the research status,fault tolerance control and adaptive dynamic planning of the modular manipulator.(2)Based on the modular properties of the modular manipulator,the dynamic model is derived by the Newton-Ola iteration algorithm,and the modular mechanical manipulator dynamics representation of sensor fault and actuator fault is given respectively.(3)For the modular manipulator system of multi-sensor fault,the expansion system is obtained by differential homogenization principle and combined with the first-order filter,and a design method for the optimal fault-tolerant controller of the modular manipulator with multi-fault concurrent is proposed based on the fault observer of ADP theory and design.Based on the fault observer estimation,the sensor function design performance index function,the fault tolerance control problem into the optimal control problem,the design control rate is composed of nominal control rate and approximate optimal control rate.In addition,the general approximation of neural network is combined with the approximation of the optimal performance index,and the online iterative learning is carried out by the evaluation network structure of the single network structure to obtain the optimal fault tolerance control rate.(4)For the modular manipulator system with sensor and actuator fault,the first-order filter device is designed to convert the sensor fault into a pseudo-actuator failure,and the performance index function is designed in combination with the observation of adaptive faulter,so as to obtain the HJB equation.Combined with ADP theoretical knowledge,the expression form of the optimal fault tolerance control rate of the modular manipulator is obtained,and on this basis,the knowledge of the neural network and the on-line iterative algorithm of the design,the optimal fault tolerance control rate of our design is obtained,and finally,the validity of the proposed algorithm is further proved by simulation experiment.