Adaptive Dynamic Programming for Finite-Horizon Networked Control System with Packet Loss and Delay

This paper presents a novel approach to adaptive optimal control of networked control systems with packet loss and delay using adaptive dynamic programming. The impact of stochastic packet loss and delay is first considered. The relationship between the stochastic Riccati equation and the deterministic Riccati equations is established via simulations that result from sampling the original system. It enables the implementation of stochastic approximate optimal control by means of an off-policy iterative strategy. The effectiveness of proposed procedures is finally demonstrated through an example.