optimal control of unknown continuous time linear periodic systems by adaptive dynamic programming with applications to magnetic attitude control

Abstract This article studies the data‐driven controller design of continuous‐time linear periodic systems. By treating the optimal control problem as a two‐player zero‐sum game problem, the adaptive dynamic programming technique is utilized. Based on the Newton's method, a one‐loop model‐based policy‐iteration (PI) method is introduced. Since the initial matrix for starting the proposed model‐based PI method is not easy to choose, a data‐driven value iteration method without requiring the knowledge of the initial matrix is then proposed. The convergence of the proposed methods is analyzed and the numerical simulation verifies the effectiveness of the proposed scheme.