Research On Intelligent Control Of Electrically Driven Robotic Arm Considering Joint Friction And Joint Flexibility

In the past few years,researchers around the world have been doing a lot of research work in the development of robotic manipulator controllers due to the large amount of early accumulation in expanding scientific knowledge and exploring frontier fields.As a result,robotic manipulators have been used in a large number of applications in various fields,especially in space exploration,medical rehabilitation,industrial manufacturing,etc.Generally speaking,robotic manipulators are complex,highly coupled nonlinear systems with various uncertainties that cannot be modeled,so it is impractical to model the robot manipulators in minute detail.It has become a research hotspot in the field of manipulator control to design a controller with strong robustness,high control accuracy and fast response for the manipulator system with model uncertainty and timevarying external disturbance.In recent years,emerging intelligent control methods can approximate unknown models and break the dependence of controller design on model parameters.Therefore,this paper considers the combination of traditional model-based control method and modern intelligent control technology,and always designs a control method with excellent performance based on the two aspects of robotic manipulator model optimization and the use of new intelligent control technologies.Being a conventional control method,sliding mode control is extensively employed in the domain of control of nonlinear systems because of its simple design method and strong robustness.Especially after several generations of research,the dynamic sliding mode surface,terminal sliding mode surface,fast terminal sliding mode surface and other sliding mode control methods with faster convergence and stronger robustness have been developed.However,as a model-based control method,its control effect depends on the accuracy of the controlled model to some extent.As an external force affecting the motion of the manipulator,joint friction should be modeled more accurately.Therefore,in this paper,the LuGre friction model is first introduced to optimize the manipulator model,and then an error-driven non-singular fast terminal sliding mode surface is used to design the controller based on this model.To solve the chattering problem inherently in sliding mode control,an enhanced power reaching law is designed.For the unmodeled and unknown portions of the system,a model-free control scheme is applied to reduce the dependence of the controller design on the model,and the delay estimation is used to estimate the unmodeled part.In addition,in order to improve the transient performance of the system,the prescribed performance function is applied to confine the tracking error of the robotic manipulator.The simulation performance indicates that our controller is superior to other advanced control methods.Based on the above research results,the robotic manipulator model is further optimized.The dynamic model of the electrically driven flexible joint manipulator(electrically driven flexible joint robotic manipulator,EDFJRM)was established by taking into account the flexible phenomenon of joint movement and the circuit model of the driving motor.For this model,backstepping control is used as the basic framework for designing the controller.At the same time,radial basis function neural network(radial basis function neural network,RBFNN)is employed to approach the uncertain model of the system.To solve the differential explosion problem inherent in backstepping control,a second-order command filter is introduced and an error compensation mechanism(error compensation mechanism,ECM)is developed to provide compensation for filtering errors.To resolve this issue of mismatched disturbance in the newly established model,an estimator expressed in matrix form is designed to directly compensate for the problem.Meanwhile,it also makes up for the shortcomings of the neural network estimation error which cannot converge asymptotically and responds slowly to the external time-varying disturbance.Finally,the benefits of our controller are proved by comparing the proposed control method with several advanced control methods through simulation experiments.