Linear structures in nonlinear optimal control

TL;DR

This paper explores optimal control of affine dynamical systems with nonlinear states, deriving approximate solutions via singular perturbation methods and identifying conditions for exact solutions in specific classes.

Contribution

It introduces a novel singular perturbation approach to affine optimal control problems, providing exact solutions for certain nonlinear trajectory tracking cases.

Findings

Approximate solutions for small regularization parameter  derived

Conditions for linear evolution equations identified

Exact solutions obtained for specific nonlinear systems

Abstract

We investigate optimal control of dynamical systems which are affine, i.e., linear in control, but nonlinear in state. The control task is to enforce the system state to follow a prescribed desired trajectory as closely as possible, a task also known as optimal trajectory tracking. To obtain well-behaved solutions to optimal control, a regularization term with coefficient $\varepsilon$ must be included in the cost functional. Assuming $\varepsilon$ to be small, we reinterpret affine optimal control problems as singularly perturbed differential equations. Performing a singular perturbation expansion, approximations for the optimal tracking of arbitrary desired trajectories are derived. For $\varepsilon=0$, the state trajectory may become discontinuous, and the control may diverge. On the other hand, the analytical treatment becomes exact. We identify the conditions leading to linear evolution equations. These result in exact analytical solutions for an entire class of nonlinear trajectory tracking problems. The class comprises, among others, mechanical control systems in one spatial dimension and the FitzHugh-Nagumo model with a control acting on the activator.