Adversarial Tracking Control via Strongly Adaptive Online Learning with Memory

TL;DR

This paper introduces a new adversarial tracking control method that leverages strongly adaptive online learning with memory, providing robust performance guarantees against adversarial disturbances in linear systems.

Contribution

It develops a comparator-adaptive algorithm for online linear optimization with movement cost and a novel strongly adaptive algorithm with memory, connecting these to adversarial tracking control.

Findings

Achieves near-optimal performance without tuning in online linear optimization.

Introduces the first reduction from adversarial tracking control to strongly adaptive online learning.

Provides strong guarantees for tracking large-range reference trajectories.

Abstract

We consider the problem of tracking an adversarial state sequence in a linear dynamical system subject to adversarial disturbances and loss functions, generalizing earlier settings in the literature. To this end, we develop three techniques, each of independent interest. First, we propose a comparator-adaptive algorithm for online linear optimization with movement cost. Without tuning, it nearly matches the performance of the optimally tuned gradient descent in hindsight. Next, considering a related problem called online learning with memory, we construct a novel strongly adaptive algorithm that uses our first contribution as a building block. Finally, we present the first reduction from adversarial tracking control to strongly adaptive online learning with memory. Summarizing these individual techniques, we obtain an adversarial tracking controller with a strong performance guarantee even when the reference trajectory has a large range of movement.