Data-Driven Output Feedback Control for Linear Time-Delay Systems

This paper proposes an adaptive learning-based output feedback optimal control (OFOC) algorithm for linear time-delayed systems, aiming to address control issues in systems with delays. Time-delay systems are prevalent in various engineering fields, yet traditional control methods typically rely on complete system state information. To overcome this limitation, a novel adaptive control algorithm is introduced, which estimates the optimal feedback gain using only system output information. Additionally, drawing from the concept of zero-sum games, an online learning algorithm is developed that can update the feedback gain in real-time without requiring full state information, ensuring both system stability and optimal performance. By constructing an error-driven learning process, system parameters are dynamically estimated, eliminating the need for precise state measurements. Simulation results demonstrate that the proposed method exhibits superior control performance and robustness in practical delayed systems.