Adaptive SARs for Physical Recovery

PhD Project - Learning and adapting socially assistive robot behaviour during long-term upper-limb recovery using reinforcement learning.

Socially Assistive Robots (SARs) have demonstrated strong potential in rehabilitative scenarios due to their positive effect on user engagement and suitability for tasks traditionally undertaken by expert therapists. For SARs supporting recovery from upper-limb injuries, user personalization is key to sustaining engagement over the long term. This project attempts to utilize participatory design with expert physiotherapists to develop and train, through interactive reinforcement learning, an SAR companion which adapts its behaviour in real time to the needs and preferences of a person recovering from upper-limb injury.


Carl Bettosi

PhD Researcher

Edinburgh Centre for Robotics

Prof. Lynne Baillie


Department of Computer Science (MACS), Heriot-Watt University

Dr Susan Shenkin

2nd Supervisor

Department of Clinical and Surgical Sciences, The University of Edinburgh

Dr Christian Dondrup

3rd Supervisor

Department of Computer Science (MACS), Heriot-Watt University


Project funded by the Engineering and Physical Sciences Research Council (EPSRC).

Date: September 2021 - September 2025

EPSRC reference: EP/S023208/1