Highly efficient robot dynamics learning by decomposed connectionist feedforward control structure

A major objective in this paper is the application of connectionist architectures for fast and robust online learning of dynamic relations used in robot control at the executive hierarchical level. The proposed connectionist robot controllers as new feature use decomposition of robot dynamics in the space of internal robot coordinates. In this way, this method enables the training of neural networks on the simpler input-output relations with significant reduction of learning time. The proposed controller structure comprises a form of intelligent feedforward control in the frame of decentralized control algorithm with feedback-error or driving torque error learning method. The another important features of these new algorithms are fast and robust convergence properties because the problem of adjusting the weights of internal hidden units is considered as a problem of estimating parameters by recursive least square method and Kalman filter approach. From simulation examples of robot trajectory tracking it is shown that when a sufficiently trained network is desired, the learning speed of the proposed algorithms is faster than that of the standard back propagation algorithm.