Nonlinear Control of Partially Known Systems Based on Polynomial Representation and Reinforcement Learning
Yuta Tsuge, Tanagorn Jennawasin, Tatsuo Narikiyo, Michihiro Kawanishi · IEEJ Transactions on Electronics Information and Systems · 2015
In this study we propose a stabilizing controller for partially known systems.We assume that the partially known systems can be identified as polynomial systems on a neighborhood of equilibrium point.For polynomial systems we give a stabilizing controller and the domain of attraction(DOA) by means of the Lyapunov function.Since the reinforcement learning is carried out by the Lyapunov function, the reinforcement learning controller can be connected to the stabilizing controller of the polynomial systems on the DOA.To demonstrate the usefulness and validity of the proposed method, we apply the proposed control method to the Furuta pendulum.