Monday, 24 February 2014

Wearable Computers - A cure for Sleep Paralysis?

This is a paralysis that occurs on waking from sleep, usually in the early morning in some people.
Here is an introduction to sleep paralysis,  report  from the NHS.   The mind works, the person can control breathing and possibly also foot movement, but the rest of the body is paralysed. The sensation may only last a few minutes but is terrifying, Drugs can be used to cure it but a non invasive method may be preferred.  Sleep paralysis  also seems to run in families and I suffer from it at least once a week. Every time I have it I am determined to find a fix or cure for it, but on waking and the morning arriving,  the fear has gone.
 I now  have time for revisiting  sleep paralysis. I also had previous not put much effort into it as I thought of interest to only a few people but  it may occur in up to 8% of the population.
The proposed  sensor device is attached to  the foot close to the toes under a band aid  (the person can wriggle toes in sleep paralysis).   A bio sensor detects motion and wakes up the person fully by a piezo  alarm. The sensor will be washable and reusable and should cost around £20.  The software algorithms require the right degree of sensing so every time some one moves in bed they are not woken up, only when in sleep paralysis.  Breathing is easier to detect but quite invasive to go to sleep every night with a chest  monitor (I did this once for 4 weeks but not again).

There has been prior work on using sensors for detecting sleep paralysis but the devices were quite cumbersome for wearing every night.  I also build a  proof of concept for  an Early Onset Epilepsy detection system for the p3i  Lab of Northumbria University in 2011.
Here are some earlier notes I did on sleep paralysis in 2012. I proposed an ankle chain but in the last two years we now have a smaller  ARM computer 3mm x 3mm x 1m with built in accelerometer, and tiny Lithium Polymer batteries so can fit under a band aid.

ST Microcontroller 

Most of the research work will be in the software for optimising the real time filtering algorithms and user trials.  The hardware proof of concept can be tested with the Texas SensorTag.

Contact Lyndsay Williams or +44 (0) 7970 101578.