2020/12/08 by Alexandros Tanzanakis, John Lygeros, Tanzanakis, Alexandros +1
Computer Science · Engineering · #Adaptive Dynamic Programming Control #Advanced Control Systems Optimization #FOS: Electrical engineering #Mechanical Circulatory Support Devices #Systems and Control (eess.SY) #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2012.04318
openalex publication_date 2020/12/08 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
We study the problem of optimal state-feedback tracking control for unknown\ndiscrete-time deterministic systems with input constraints. To handle input\nconstraints, state-of-art methods utilize a certain nonquadratic stage cost\nfunction, which is sometimes limiting real systems. Furthermore, it is well\nknown that Policy Iteration (PI) and Value Iteration (VI), two widely used\nalgorithms in data-driven control, offer complementary strengths and\nweaknesses. In this work, a two-step transformation is employed, which converts\nthe constrained-input optimal tracking problem to an unconstrained augmented\noptimal regulation problem, and allows the consideration of general stage cost\nfunctions. Then, a novel multi-step VI algorithm based on Q-learning and linear\nprogramming is derived. The proposed algorithm improves the convergence speed\nof VI, avoids the requirement for an initial stabilizing control policy of PI,\nand computes a constrained optimal feedback controller without the knowledge of\na system model and stage cost function. Simulation studies demonstrate the\nreliability and performance of the proposed approach.\n