SMARTAR: Augmented Reality platform for increasing mobility and independence of Parkinson's patients
Background
Parkinson's disease is a progressive neurological condition that affects over 6 million people worldwide. Half of Parkinson patients suffer from a condition called Freezing of Gait (FOG), where they feel as if their feet are "glued" to the ground. This sensation can occur when you start to walk or while walking and may last for several seconds to minutes. In the UK there are 145,000 people with Parkinson's and over 72,000 people who suffer from Freezing of Gait. FOG not only contributes to falls and related injuries, but also compromises quality of life as people often avoid engaging in functional daily activities both inside and outside the home.
Objective
This project focused on developing SMARTAR -- Augmented Reality platform for increasing the mobility and independence of Parkinson's disease patients. SMARTAR is an Augmented Reality Glasses that, through the use of sensors, will monitor a person's gait, detecting if a freezing incident occurs. It will then use proven techniques of giving the user a visual focus point of parallel lines on the ground to "step over".
Benefits
This method has been proven to be the best solution to overcoming Freezing. SMARTAR is a portable solution that works both inside and outside, allowing the user to keep their mobility and to be more independent and less reliant on family members or caretakers. Other solutions addressing freezing in Parkinson patients require the user to turn them on/off or to always be on. They can also be very noticeable, drawing unwanted attention to the person with Parkinson's.
Brunel Innovation Centre's Role
The main challenge addressed by Brunel University was to develop a suite of methodologies and algorithms for gait characterisation in Parkinson's patients. Brunel has successfully delivered:
- Deep learning models with variable computational requirements for maximum compatibility with multiple devices for gait characterisation
- Deep learning models with variable computational requirements for maximum compatibility with multiple devices for action characterisation (walking, standing, sitting etc)
- A novel methodology for automatic retraining to allow fast and easy calibration to new patients by using 5 minutes of data for a new patient. Gait analysis algorithms have been developed in the past using similar sensors, but those sensors were attached to arms, legs and waist, making them invasive and disruptive to everyday life.
Project Partners
EMTEQ Ltd
The Imagination Factory
Brunel University London