Learning from Adaptive Neural Output Feedback Control of Robot Manipulators

Abstract

Abstract This paper studies the problem of learning from adaptive neural network (NN) control of a one-link robot manipulator including motor dynamics in uncertain dynamical environments. With the employment of a newly state transformation and a high-gain observer, the one-link robot system is transformed into a norm form, and then only one NN is employed to approximate the lumped uncertain system nonlinearity in the adaptive control design. Partial persistent excitation (PE) condition of radial basis function (RBF) NNs is satisfied during tracking control to a recurrent reference trajectory. Under the PE condition, the proposed adaptive NN control is shown to be capable of acquiring knowledge on the uncertain robot dynamics in the stable control process and of storing the learned knowledge in memory. Subsequently, a novel neural learning control technique exploiting the learned knowledge without readapting to the unknown robot dynamics is developed to achieve closed-loop stability and improved control performance. Simulation studies are performed to demonstrate the effectiveness of the proposed control technique.