Data-Based H ∞ Tracking Control for Time-Delay Systems Via Adaptive Dynamic Programming

In this paper, an data-based H ∞ tracking control algorithm is studied for the discrete delay system with disturbances. The augmented system is constructed and then transformed into a data-based system consist of inputs, outputs, and reference trajectories, which replace the state vector by a data vector. Next, based on the Bellman optimality principle, the historical data is adopted to derive a new Bellman equation. In order to realize H ∞ tracking control, the input and disturbance are generalized to the zero-sum game problem. Finally, by collecting operational data online, a H ∞ tracking control method based on databased reinforcement learning algorithm is presented. Simulation example shows that the proposed reinforcement learning tracking control method is effective.