A Structured Online Learning Approach to Nonlinear Tracking with Unknown Dynamics

One of the most common problems in the control of dynamical systems is to track a desired reference trajectory, which is found in a variety of real‐world applications. This chapter extends the Structured Online Learning (SOL) framework to tracking with unknown dynamics. Similar to regulation problems, the applications of tracking control can benefit from Model‐based Reinforcement Learning (MBRL) that can handle the parameter updates more efficiently. It proposes an approximate optimal tracking control framework based on a particular structure of nonlinear dynamics, where a linear quadratic discounted cost is assumed. The chapter provides the details of implementation of the obtained framework as a learning‐based approach. Initially, the SOL approach was proposed for solving stabilization and regulation problems. The chapter reviews the model‐based learning framework. It focuses on the properties of the proposed control scheme rather than the identification process within the simulations.