Real-time inverse optimal control for a planar robot

Abstract

This paper presents an inverse optimal neural controller, which is constituted by the combination of two techniques: a) inverse optimal control to avoid solving the Hamilton Jacobi Bellman (HJB) equation associated to nonlinear system optimal control, and b) an on-line neural identifier, which uses a recurrent neural network, trained with the extended Kalman filter (EKF), in order to build a model of an assumed unknown nonlinear system. The applicability of the proposed approach is illustrated by real-time control of a planar robot.