Self‐learning‐based optimal tracking control of an unmanned surface vehicle with pose and velocity constraints

Abstract

AbstractIn this article, subject to both pose and velocity constraints within narrow waters, a self‐learning‐based optimal tracking control (SLOTC) scheme is innovatively created for an unmanned surface vehicle (USV) by deploying actor‐critic reinforcement learning (RL) mechanism and backstepping technique. To be specific, the barrier Lyapunov function (BLF) is devised to uniformly limit the states within a predefined region pertaining to a smoothly feasible reference trajectory. By virtue of a constrained Hamilton–Jacobi–Bellman (HJB) function, an actor‐critic control structure under backstepping is established by employing adaptive neural network identifiers which recursively updates actor and critic, simultaneously. Eventually, theoretical analysis proves that the entire SLOTC scheme can render all the states remain in the predefined compact set while tracking errors converge to an arbitrarily small neighborhood of the origin. Simulation results on a prototype USV demonstrate remarkable effectiveness and superiority.