Dynamical System Segmentation for Information Measures in Motion

TL;DR

This paper introduces a novel dynamical system segmentation algorithm to decode task-specific information from motion data, enabling detailed performance assessment and demonstrating that assistance improves task embodiment in humans.

Contribution

The paper presents a new algorithm for motion segmentation that extracts behavior alphabets without prior knowledge, advancing understanding of task embodiment in motion analysis.

Findings

Assistance improves task embodiment in human motion.

Task embodiment predicts assistance effectiveness better than mean-squared-error.

The method accurately identifies task-relevant behaviors in motion data.

Abstract

Motions carry information about the underlying task being executed. Previous work in human motion analysis suggests that complex motions may result from the composition of fundamental submovements called movemes. The existence of finite structure in motion motivates information-theoretic approaches to motion analysis and robotic assistance. We define task embodiment as the amount of task information encoded in an agent's motions. By decoding task-specific information embedded in motion, we can use task embodiment to create detailed performance assessments. We extract an alphabet of behaviors comprising a motion without \textit{a priori} knowledge using a novel algorithm, which we call dynamical system segmentation. For a given task, we specify an optimal agent, and compute an alphabet of behaviors representative of the task. We identify these behaviors in data from agent executions, and compare their relative frequencies against that of the optimal agent using the Kullback-Leibler divergence. We validate this approach using a dataset of human subjects (n=53) performing a dynamic task, and under this measure find that individuals receiving assistance better embody the task. Moreover, we find that task embodiment is a better predictor of assistance than integrated mean-squared-error.