The challenge

Many patients receive their physical rehabilitation exercises at home. However, due to lack of healthcare manpower to monitor daily exercises, therapist do not receive enough insightful feedback to be able to adjust the patient’s therapy optimally.

The proposed solution

This project aims to build a movement monitoring system that records movement and flags parts of the exercise that should be reviewed by a therapist. This will give the therapist insights into the patient’s movements and reduce time spent on feedback generation.

The project is composed of two main parts:

• A home-based, upper-limb sensing system – built on a combination of a depth camera, a pair of IMU wristbands and a suite of sensorised household items to capture the full degree-of-freedom of upper-extremity movements.
• Kinematic data from the sensing system – which will be automatically analysed to flag anomalies using statistical and machine-learning techniques. These techniques will generate the norm model for each exercise by learning from exercise demonstrations from a group of healthy subjects.