Exploring the Feasibility of Generating Physical Prescriptions Using Large Language Models with Motion Tracking Inputs

Gangmin Li*, Xuming Bai

*Corresponding author for this work

Research output: Chapter in Book or Report/Conference proceedingConference Proceedingpeer-review

Abstract

This paper reports our first attempt using Large Language Models (LLMs) to generate personalised physical prescriptions using real-time motion tracking system inputs for gait analysis. We investigate the feasibility of such a system through relevance, accuracy, and potential health outcomes of the generated prescriptions compared to those created by humans. The study highlights the benefits, challenges, and future directions of integrating sensor-based real-time inputs into LLMs for generating physical prescriptions for physical treatment and medical rehabilitation.

Original languageEnglish
Title of host publicationICAC 2025 - 30th International Conference on Automation and Computing
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798331525453
DOIs
Publication statusPublished - Aug 2025
Event30th International Conference on Automation and Computing, ICAC 2025 - Loughborough, United Kingdom
Duration: 27 Aug 202529 Aug 2025

Publication series

NameICAC 2025 - 30th International Conference on Automation and Computing

Conference

Conference30th International Conference on Automation and Computing, ICAC 2025
Country/TerritoryUnited Kingdom
CityLoughborough
Period27/08/2529/08/25

Keywords

  • Large Language Models (LLMs)
  • Motion Tracking Systems
  • Physical Prescription Generation (PPG)
  • Physical Therapy
  • Rehabilitation
  • Sensor-Based Data Integration

Cite this