Low-Burden LLM-Based Preference Learning: Personalizing Assistive Robots from Natural Language Feedback for Users with Paralysis

TL;DR

This paper introduces a low-burden, LLM-based framework for personalizing assistive robots using natural language feedback, reducing user fatigue and ensuring safety for users with paralysis.

Contribution

It presents a novel offline pipeline that translates unstructured language into safe, explicit robot control policies grounded in clinical reasoning, verified by automated safety checks.

Findings

Significantly reduces user workload compared to traditional methods.

Generated policies are confirmed safe and accurately reflect user preferences.

Validated in a simulated meal preparation study with 10 adults with paralysis.

Abstract

Physically Assistive Robots (PARs) require personalized behaviors to ensure user safety and comfort. However, traditional preference learning methods, like exhaustive pairwise comparisons, cause severe physical and cognitive fatigue for users with profound motor impairments. To solve this, we propose a low-burden, offline framework that translates unstructured natural language feedback directly into deterministic robotic control policies. To safely bridge the gap between ambiguous human speech and robotic code, our pipeline uses Large Language Models (LLMs) grounded in the Occupational Therapy Practice Framework (OTPF). This clinical reasoning decodes subjective user reactions into explicit physical and psychological needs, which are then mapped into transparent decision trees. Before deployment, an automated "LLM-as-a-Judge" verifies the code's structural safety. We validated this system in a simulated meal preparation study with 10 adults with paralysis. Results show our natural language approach significantly reduces user workload compared to traditional baselines. Additionally, independent clinical experts confirmed the generated policies are safe and accurately reflect user preferences.