Predicting the consequence of action in digital control state spaces

TL;DR

This paper explores fundamental challenges in learning control laws in continuous state spaces for artificial devices, proposing an approach inspired by neuroscience principles to improve motor task learning.

Contribution

It introduces a novel control framework inspired by neuroscience, moving away from classical displacement control to end effector control for better learning in continuous spaces.

Findings

Proposes a neuroscience-inspired control principle.

Addresses limitations of classical control in learning contexts.

Lays groundwork for more effective motor learning algorithms.

Abstract

The objective of this dissertation is to shed light on some fundamental impediments in learning control laws in continuous state spaces. In particular, if one wants to build artificial devices capable to learn motor tasks the same way they learn to classify signals and images, one needs to establish control rules that do not necessitate comparisons between quantities of the surrounding space. We propose, in that context, to take inspiration from the "end effector control" principle, as suggested by neuroscience studies, as opposed to the "displacement control" principle used in the classical control theory.