Direct Data-Driven Optimal Set-Point Tracking Control of Linear Discrete-Time Systems

This brief discusses the direct data-driven optimal set-point tracking (OST) control problem of linear discrete-time systems. A cost function without discount factor is defined in terms of control increment, which is different from the popular discount cost function for OST control problems. The standard linear quadratic tracking (LQT) solution is reformulated by convex duality, and the algebraic Riccati equation (ARE) for finding optimal solution of the infinite horizon OST problem is equivalently converted to a convex optimization (CO) problem. More importantly, an OST controller which ensures the tracking error converge to zero only using off-line historical data. Finally, an example of 2-degree-of-freedom (2-DOF) helicopter verifies the effectiveness of the obtained results.