Adaptive Communication System for Individuals with Motor Neurone Disease

Motivation & Background

Motor Neurone Disease (MND) is a progressive neurological condition that progressively damages motor neurons in the brain and spinal cord leading to muscle weakness and wasting that can affect walking, talking, eating and breathing. This progressively limits a persons ability to communicate due to the loss of motor control. MND affects every individual differently, some lose their voice first, others have trouble with their arms and legs. Existing assistive communication systems are generalised and lack any personalisation. This can make daily communication frustrating and difficult. A system that leverages artificial intelligence and adaptive machine learning to personalise communication could significantly improve the quality of life for individuals with MND.

## Project

The aim of this project is to design and develop an AI powered adaptive communication system that assists people with MND to communicate. There will be 3 input options - (to include all stages of the disease)

• Microphone
• Camera - for eye gaze, MND does not affect the eyes
• On screen buttons

Machine learning will deliver personalised context aware auto completion suggestions to help the individual complete their communication more effectively. The model would adapt dynamically to the users specific routine, habits and progression of the disease. This would be based on a mobile application. Two styles of accounts would be available, an individual with MND or a family member. Families will be able to create groupchat style set ups in which they can communicate with each other. The individual with MND will have the assistive technology on their account, and when they use the system they can either play the message out loud for in person communication or send via a message for online communication. This will also keep families connected in this tough time. Further app features can be developed and implemented further down the line.

Approach

• Front end - React native (JavaScript), Tailwind CSS
• Back end - Python
• Database/cloud service - Azure/Firebase
• Machine learning - TensorFlow or PyTorch