Finite sample learning of moving targets

TL;DR

This paper extends randomized learning techniques to moving targets, providing sample complexity bounds and a MILP-based method for convex polytope targets, with application to autonomous emergency braking.

Contribution

It introduces a novel PAC estimation approach for moving targets, including a MILP method for convex polytopes, advancing learning in dynamic environments.

Findings

Derived a new sample complexity bound for moving targets.

Developed a MILP-based method for convex polytope targets.

Demonstrated application in autonomous emergency braking.

Abstract

We consider a moving target that we seek to learn from samples. Our results extend randomized techniques developed in control and optimization for a constant target to the case where the target is changing. We derive a novel bound on the number of samples that are required to construct a probably approximately correct (PAC) estimate of the target. Furthermore, when the moving target is a convex polytope, we provide a constructive method of generating the PAC estimate using a mixed integer linear program (MILP). The proposed method is demonstrated on an application to autonomous emergency braking.