Robust Adaptive Neural Network Control of Uncertain Rheonomically Constrained Manipulators

In the presence of uncertain robot model and external disturbance, the robust adaptive neural network position/force control of manipulators constrained on time- varying surfaces is developed in this paper. Firstly, a reduced dynamic model is proposed through a nonlinear transformation by some assumptions. Then, by introducing a new variable, the time-varying differential-algebraic reduced dynamic model is transformed into a nonlinear system. After that, two neural network systems are designed to approximate two nonlinear functions by virtue of the ability in approximating an arbitrary nonlinear function. And also, parameter adaptive laws and robust control terms are given to guarantee the stability of the closed-loop system. Finally, simulation results are conducted to demonstrate the effectiveness of the proposed controller.