Self-tuning control by neural networks

Abstract

A new self-tuning controller consisting of a PD controller, an inverse dynamics compensator, and a neural controller is proposed. In order to train the neural controller located in front of a system, the inverse dynamics of the system is used to calculate the inverse Jacobian of the unknown system. With the neural identifier the overall control architecture can be made stable. The control performance is compared with that of a conventional controller without the neural networks. Computer simulation results show that the proposed control architecture is effective in controlling of a robotic system.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>