Research Of Manipulator Trajectory Tracking Control Based On COA-PIDNN

The research of manipulator trajectory control is a crucial part of the manipulatorcontrol. It is significant for promoting the development of robot technology, improving theefficiency of social productive, the emancipation of the productive forces and acceleratingthe construction of new socialist construction. In this paper, the research of manipulatortrajectory tracking control based on COA-PIDNN (Proportional-Integral-DerivativeNeural Network optimized by Chaos Optimization Algorithm) is studied.Firstly, in view of the defect of conventional PIDNN, some improvements of PIDNNara made based on chaos optimization algorithm (COA). Conventional PIDNN optimizedby BP algorithm, namely BP-PIDNN, exists the shortage of easily falling into localminimum. Two improved PIDNN strategies, MSCOA-PIDNN and NMSCOA-PIDNN,are proposed by respectively using two kinds of COA, the MSCOA (Mutative ScaleChaos Optimization Algorithm) and the NMSCOA (New Mutative Scale ChaosOptimization Algorithm), to optimize the weights of PIDNN. In consideration of theproblem of manipulator which has uncertainty and strong coupling, the two improvedPIDNN were used to identify the manipulator. To verify the validity of the identificationmodel, a two degrees of freedom manipulator is used as the object of study in thesimulation research. The two improved PIDNN are compared with BP-PIDNNrespectively. The simulation results show the validity of identification model and theperformance improvements of improved PIDNN.Secondly, in consideration of the strong coupling and uncertainty of the multi-degreeof freedom manipulator, two multi-step predictive neural network inverse controlstrategies based on MSCOA-PIDNN and NMSCOA-PIDNN respectively are proposed onpremise of the identification model in this paper. The Stability proofs are given. The twodegrees of freedom manipulator is regarded as the controlled object in the controlsimulation research. The results show the validity of the control methods. The BP-PIDNNis also used to control manipulator in the simulation. The control results of BP-PIDNNcompared with the control results of MSCOA-PIDNN and NMSCOA-PIDNN respectively.The comparisons show the performance improvements of improved PIDNN ulteriorly.