Modeling And Control Of Flexible Manipulator

Based on a physical model of a two-link flexible manipulator, a study on themodeling and control of flexible manipulator, using the neural network predictivecontrol, is presented in the paper.Firstly, the dynamic equations of the flexible manipulator are developed. Thephysical model of the flexible manipulator is reduced by taking the flexible link asEuler-Bernoulli beam. The dynamic equations of the flexible manipulator aredeveloped by the Lagrange equation and the assumed mode method. And the vibrationanalysis obtained by the Newmark-β method well verifies the accuracy of the model.Secondly, the BP neural network identification model is set up. Constructing a BPneural network identification system, and using the data from the vibration analysis totrain and test it, thus we can get the BP neural network predictor.Thirdly, the trajectory of rigid flexible manipulator is planned. Using theOptimization method with the velocity of the arm as its optimization goal and thecentral difference method to plan the ideal trajectory of a two-link rigid manipulator,this is the desired trajectory of the controller.Finally, the controller of the flexible manipulator is integrated designed. The BPneural network predictor, ideal trajectory and a newly neural network controller arecombined to be a neural network predictive control system. The simulation shows thatthe flexible manipulator can move smoothly according to the desired trajectory underthe control of the neural network predictive control system.