Improving User Specifications for Robot Behavior through Active Preference Learning: Framework and Evaluation

TL;DR

This paper presents a framework for improving robot behavior specifications through active preference learning, enabling non-expert users to iteratively refine constraints and enhance robot performance in complex tasks.

Contribution

The authors introduce an iterative preference learning framework that refines user-specified constraints, improving robot task performance and consistency across different users.

Findings

Most users accept alternative solutions, enabling specification revision.

Revised specifications significantly improve robot performance.

The learning process reduces variability between user specifications.

Abstract

An important challenge in human-robot interaction (HRI) is enabling non-expert users to specify complex tasks for autonomous robots. Recently, active preference learning has been applied in HRI to interactively shape a robot's behavior. We study a framework where users specify constraints on allowable robot movements on a graphical interface, yielding a robot task specification. However, users may not be able to accurately assess the impact of such constraints on the performance of a robot. Thus, we revise the specification by iteratively presenting users with alternative solutions where some constraints might be violated, and learn about the importance of the constraints from the users' choices between these alternatives. We demonstrate our framework in a user study with a material transport task in an industrial facility. We show that nearly all users accept alternative solutions and thus obtain a revised specification through the learning process, and that the revision leads to a substantial improvement in robot performance. Further, the learning process reduces the variances between the specifications from different users and, thus, makes the specifications more similar. As a result, the users whose initial specifications had the largest impact on performance benefit the most from the interactive learning.