Research On Switching Control Of Robotic Arm System Based On Adaptive Neural Networ

In the context of the continuous development of science and technology,robots have received extensive attention as a powerful tool for industrial production.Robots can replace humans to do dangerous work in the scene or replace computationally intensive and repetitive actions.They have a pivotal position in the fields of surgical medicine,home services,space exploration and industrial production.As a key component of the industrial robot system,the manipulator is essential in robot control because it can achieve high-precision tracking of the desired trajectory.However,in industrial production,the load of the manipulator is usually constantly changing,which makes it difficult to accurately establish its dynamic system model.At the same time,it will be affected by nonlinear factors such as unknown dead zone,input saturation and output limitation,which will reduce the control performance of the manipulator when performing preset trajectory actions.Therefore,it is of great practical significance to study the trajectory tracking control problem of the manipulator system under such complex conditions.Based on RBF neural network theory,adaptive control algorithm,switching system control theory and sliding mode variable structure control method,this paper further studies the high-precision tracking control problem of uncertain manipulator dynamics system,and obtains some results in chattering suppression and control performance optimization of closedloop system under complex working conditions.The specific research contents are as follows:1.Aiming at the chattering problem in the design of sliding mode controller for robot system,a controller design scheme with variable sliding mode gain is studied.Based on the traditional sliding mode controller design,the innovation of the control scheme is that the switching gain of the designed controller can realize dynamic adaptive adjustment,that is,based on RBF neural network,the switching gain changes dynamically with the joint parameters to adapt to the unmodeled dynamics and unknown disturbances of the system.By adding an appropriate adaptive control algorithm,the lumped disturbance is effectively suppressed.Moreover,the Lyapunov method is used to prove that the trajectory tracking error of the system asymptotically converges to zero.Finally,the simulation results show that the proposed scheme reduces the system chattering and effectively improves the tracking accuracy.2.For the uncertain rigid manipulator system with variable load,a neural network switching tracking control problem based on average dwell time is studied.The scheme regards the rigid manipulator system with different loads as a switching system,and designs controllers for each subsystem based on the average dwell time principle.In each subsystem,RBF neural network is used to approximate the system structure parameters to avoid the dependence of the controller on the accurate model of the system.At the same time,a robust compensation term based on neural network is designed to suppress the influence of lumped disturbance on the system.Then,the uniform ultimate boundedness of the tracking error is proved by using the multiple Lyapunov function method.Finally,the simulation results show that the proposed control scheme can not only achieve high-precision tracking of the desired trajectory of the variable load manipulator,but also effectively weaken the chattering of the input torque.3.For the uncertain rigid manipulator system with variable loads,a new adaptive neural network switching control scheme is further proposed.Based on an improved average dwell time method,the design algorithm of adaptive output feedback neural network switching controller in result 2 is optimized.Then,based on the multiple Lyapunov function method,the boundedness of the error signal under the action of a specific switching law is proved.Finally,the simulation results verify the rationality of the design scheme.4.As for the tracking control problem of robot system with unknown dead zone,an adaptive switching controller method based on dead zone compensation is studied.Firstly,in order to address the adverse effects of load changes on the system,the switching scheme in result 2 is used in this study.For the unknown dead zone of the system,two mutually coupled neural networks are used to observe and compensate it,and a neural network-based robust compensation term is designed to eliminate potential disturbances and estimation errors.Then,the boundedness of the error signal is analyzed by using the average dwell time method,the multiple Lyapunov function method and the adaptive control law.Finally,the simulation results verify the effectiveness of the proposed control scheme.