A Review of Unmanned Vehicle Control with Adaptive Dynamic Programming Implementations

Abstract

In this review, the optimal control designs via adaptive dynamic programming (ADP) of unmanned vehicles are investigated. Various complex tasks in unmanned systems are addressed as fundamental optimal regulation and tracking control problems related to the position and attitude of vehicles. The optimal control can be obtained by solving the Hamilton-Jacobi-Bellman equation using ADP-based control methods. Neural network implementations and policy iterative ADP algorithms are common approaches in ADP-based control methods, enabling online updates and partially model-free control for unmanned vehicles with various structures. For complexities and uncertain disturbances in unmanned vehicle dynamics, robust ADP-based control methods are proposed, including robust ADP control for matched and unmatched uncertainties, robust guaranteed cost control with ADP, and ADP-based $H_{\infty }$ control. In order to reduce communication and computational costs in unmanned vehicle operations, a preliminary discussion on event-triggered optimal control using ADP-based control methods is presented.