Switching Adaptive Control For A Class Of Robot Systems Based On Average Dwell Time

Industrial manipulators have been widely used in applications.With the rapid development of automation of manufacturing industry,industrial manipulators have begun to appear in every field of different industrial situations.One of the core problems is how to control the mechanical arm model with higher accuracy and less energy consumption.At the same time,how to solve the problem of parameter variation and influence of friction in the manipulator is of considerable practical significance as well.This thesis studies adaptive switching tracking control for mechanical manipulators with parameter uncertainties.Applying the switching control technique,we design adaptive tracking controllers.The research results we obtain are as follows:First of all,for mechanical manipulator control systems containing parameter uncertainties and even parameter jumps,considering the actual application situations in which acceleration information is difficult to measure and the inversion of the changing matrix is time-consuming,we adopt adaptive control methods to control the robot arm.To solve the problem of parameter jumping in application,we use switching method to model the system,based on which we use a kind of switched adaptive control method without acceleration feedback and time-varying matrix inversion.By Lyapunov method,boundedness of the tracking error is guaranteed.Secondly,in view of the mechanical arm friction disturbance problems,we use RBF neural network to present an adaptive switching control method.The best weights of the neural network for manipulator system are estimated online,and the jump condition for parameters is handled with switching control mechanism.Similarly,we can get the boundedness of tracking error and the ideal control effect.Moreover we discuss the influence of the switching control gain on the robot control system.Applying this method,we also get the ideal control results.Finally,based on two degree freedom mechanical arm,we carry on the MATLAB simulation of our control method.The simulation results show that the proposed algorithm works quite well for mechanical arm with parameter uncertainties and friction disturbance.In the end,we summarize the research in this thesis,and discuss some relative issues that deserve further study.