AI-Driven Rehabilitation Technologies

Background and goals

Finland’s ageing population, increasing rehabilitation needs, and healthcare workforce shortages create an urgent need for new, efficient and human-centred digital solutions. Rehabilitation services are also challenged by administrative burden, fragmented care pathways, and difficulties in assessing pain, especially among children with disabilities and adults with persistent pain. Recent advances in artificial intelligence, large language models, computer vision, humanoid robotics and digital fabrication offer new opportunities to improve rehabilitation quality, accessibility and efficiency.

The goal of this interdisciplinary project is to develop and evaluate AI-driven rehabilitation technologies that support both clients and healthcare professionals. The project focuses on four interconnected areas: AI-assisted patient interviews and documentation, humanoid robot-assisted rehabilitation for children with disabilities, computer vision-based movement and pain analysis, and the feasibility of digital scanning and additive manufacturing for personalized orthoses.

Objectives and benefits

The project aims to create and assess practical AI-based tools that can support rehabilitation in real clinical and educational contexts. Key objectives include developing an LLM-based interview and documentation prototype for pain assessment, exploring how QTRobot can support pain communication among children with cerebral palsy, testing computer vision-based markerless movement analysis for functional assessment, and reviewing the feasibility of home-based digital scanning for future orthosis fabrication.

The expected benefits include improved patient-clinician communication, more structured and efficient documentation, better opportunities for children to express pain, and more objective movement assessment. The project also strengthens interdisciplinary competence among students and professionals by combining engineering, physiotherapy, occupational therapy and health technology education.

Results

The project is expected to produce several applied research outputs and prototypes. These include an AI-assisted interview and documentation prototype, a structured simulated pain interview dataset, clinical summary templates, and evaluation tools for testing how reliably AI can generate clinically useful pain summaries. The project will also develop and evaluate humanoid robot-based rehabilitation applications and computer vision-based tools for movement analysis and early identification of pain-related movement patterns.

Further results include a feasibility review of home-based scanning and 3D-printing approaches for personalized orthoses, peer-reviewed publications, conference presentations, student theses, and integration of findings into Arcada’s teaching. The project will also contribute to the long-term academic development of the research group, including doctoral research.

Societal impact

The project addresses major societal challenges related to chronic pain, disability, healthcare accessibility, and workforce sustainability. By reducing documentation burden and supporting more efficient clinical workflows, AI-based tools may release healthcare professionals’ time for patient interaction and improve the quality of care. For clients, the project aims to support more continuous, personalized and accessible rehabilitation services.

The project also contributes to Finland’s competence in trustworthy AI and digital health technologies. By educating future professionals who understand both healthcare needs and technical systems, the project supports long-term innovation in rehabilitation. Its focus on children with disabilities and adults living with pain gives the work strong relevance for health equity, participation and quality of life.

Abstract

This interdisciplinary project develops and evaluates AI-driven rehabilitation technologies to improve care quality, accessibility and clinical efficiency. It responds to growing rehabilitation needs, healthcare workforce shortages and the burden of chronic pain by combining large language models, computer vision, humanoid robotics and digital fabrication. The project focuses on four areas: AI-assisted pain interviews and documentation, QTRobot-supported rehabilitation for children with disabilities, markerless movement and pain analysis, and a feasibility review of digital scanning for personalized orthoses. Expected outcomes include prototypes, simulated datasets, evaluation tools, publications, theses and integration into education. The project strengthens trustworthy AI competence and supports more human-centred, personalized and sustainable rehabilitation services.