Iterative Learning-based Trajectory Tracking Control Algorithm For Power Maintenance Robotic Arm

With the development of science and technology,it has become common to use robotic arms to accomplish related tasks,in order to realize the maintenance work,in the case of energized transmission lines.Robotic arm systems are a typical class of dynamical systems,that are highly nonlinear,strongly coupled,and have significant time variability.Accurate modeling of the dynamics of the robotic arm system is very difficult,therefore,designing a control algorithm that is quite adaptable and easy to use in the robotic arm trajectory tracking problem is of great research significance and practical value for the robotic arm to achieve real-time tracking of the desired trajectory.Iterative learning control in the mathematical sense is a control algorithm,that can be strictly defined to solve the difficult modeling of robotic arm systems,as well as the control problems of nonlinear and strongly coupled objects,and the specific research of the thesis is as follows.(1)In view of the common telescopic structure of the robotic arm carried by the transmission line inspection robot,a modeling method based on the Lagrangian method for the telescopic robotic arm linkage is proposed,which effectively improves the accuracy of the robotic arm modeling.(2)To solve the convergence problem of robotic arm trajectory tracking,an exponential iterative learning control method based on error variable gain is proposed under the condition that the initial state is not shifted.The simulation results show that,the proposed algorithm achieves better convergence speed,and sensitivity to small errors.(3)For the first type of system initial value problem(the initial value is determined but does not lie on the desired trajectory),an iterative learning control algorithm based on state gain is proposed,which uses the change of state variables between two iterations to constitute the correction quantity;for the second type of system initial value problem(the initial value does not lie on the desired trajectory and is not regular),an iterative algorithm based on initial state learning is proposed.learning control algorithm,which learns the system state variables separately.The above two algorithms solve the problem that conventional algorithms cannot converge effectively under two types of initial value conditions,and the simulation results show that,the algorithm proposed in thesis has a better suppression effect on the initial state error of the robotic arm and has better stability.Based on iterative learning control,thesis first optimizes the modeling for the scalable structure,after that,accelerates the convergence process of the algorithm under the condition that the initial value does not shift,and finally,proposes a control algorithm suitable for robotic arm trajectory tracking under the condition that the two initial states shift to solve the algorithm convergence problem.The experimental results show that the algorithm proposed in thesis achieves good results under the above conditions,and can provide some references for the trajectory tracking problem of the electric power inspection robotic arm.