Trajectory Tracking Control Research On Coordinated Manipulator Gripping Under Multi Unknown Environment

Multi-coordinated manipulator system is more and more used in various industrial applications because of its greater carrying capacity,higher flexibility,and higher production efficiency.However,the coordinated manipulators will form a closed-loop kinematic chain during the coordinated gripping tracking process,with more kinematic constraints such as position,velocity,etc.,which may lead to unnecessary internal forces or even damage to the system.Meanwhile,in the face of increasingly complex industrial application environment,multi-manipulator coordination will encounter more complex unknown interference,unmodeled dynamics and other issues direct impacting on the control accuracy.Therefore,trajectory tracking control research on coordinated manipulator gripping under multi unknown environment has a wide application prospect and important research value.Aiming at the common problems of multi coordinated manipulator system,this paper studies and proposes a series of new universal multi coordinated manipulator control methods.This paper provides a new idea for the related control theory and method research.Specific research work is as follows:(1)Construct kinematic and dynamic model of coordinated manipulator gripping under multi unknown environment.The kinematic model of the gripping system is constructed by analyzing the manipulators-workpiece closed-loop kinematic chain.Based on kinematic model,the non-squeezing load distribution model is constructed.By Lagrangian equation,the dynamic equation of single manipulator and the workpiece is constructed.Based on which,the dynamic models of the multi-manipulator gripping system under structured and unstructured unknown environment are constructed in both joint space and task space respectively.Provide the basis for control architecture design of the control method in the subsequent chapters.(2)An weighted dynamic model based adaptive neural network backstepping collaborative trajectory tracking control algorithm is proposed for the unknown environment of system dynamics parameters,as well as to dynamic compensation,trajectory tracking and weight adjustment.Based on the dynamic model of coordinated manipulator gripping system described in joint space,the multi-stage Lyapunov equation is constructed by backstepping strategy.Using a radial basis function neural network approximating the unknown part of the system dynamics,a nominal dynamic model based collaborative adaptive neural network backstepping controller is designed.The contribution between neural approximation and the nominal dynamic part are analyzed.Based on the actual situation of dynamic information,a weighted dynamic model based collaborative adaptive neural network backstepping control architecture is proposed to ensure the robustness of the system.The effectiveness of the method is proved by comparative simulation experiment and analysis.(3)An adaptive neural network synchronized internal force trajectory tracking control algorithm is proposed for the system with the unknown environment of temporary cooperation,aiming at dynamic compensation,trajectory tracking,internal force adjustment,and temporary cooperation.Virtual synchronization error is proposed to describe the relative relationship between the tracking errors of the coordinated manipulators.Combining the neural network for global dynamic approximation,the joint space controller is designed,and to achieve internal force regulation.To cope with unknown environment with unknown external load,an adaptive internal force observing synchronized sliding mode trajectory tracking control algorithm is proposed,also to deal with dynamic compensation,trajectory tracking,internal force observation,and variable load gripping.Based on the neural dynamic compensation strategy and the neural internal force observer design,the task space sliding mode controller is designed by using the synchronized coupling error to ensure the high precision tracking in the case of constant,change and other unknown external load.The stability is proved by the Lyapunov stability analysis and comparative simulation.(4)An frequency coordinated adaptive neural robust trajectory tracking control method is proposed to online calibrate the displacement parameters of multi-manipulators'bases,under inaccurate base frame parameters situation,with the goals of dynamic compensation,trajectory tracking,internal force control,and base-frame calibration.The kinematic correction model is constructed by converting the displacement parameters of the base into the kinematic parameters of the manipulator.Combine the adaptive strategy for frequency coordinated online adjustment of the modified kinematic parameters.The global structured and unstructured uncertain factors of the system are approximated by one adaptive neural network.Specialized robust techniques are set up to compensate for the instability caused by the double approximation strategies.The stability and performance of the method are proved by the Lyapunov analysis and comparative simulations,that online precision calibration capability can be obtained while internal force is kept within an acceptable range.(5)In order to realize the application of the intelligent trajectory tracking control algorithms,aiming at efficient control,control realization,a hierarchical trajectory tracking control architecture is established to solve the complicated problems caused by the above-mentioned multi unknown environment.The torque control method of general industrial manipulator based on CANopen over EtherCAT is studied.The feasibility of the method for controlling the torque mode of industrial manipulator is proved by single joint experiment,which provides the basis for the future coordinated manipulator tracking control experiment and industrial application.