Regaining a little independence

I am currently helping one of my MND friends, Patrick, to make an eyegaze controlled wheelchair driving system, or as Patrick has branded it, EyedrivOmatic. There are  a couple of systems already available but with several disadvantages including high cost and incompatibilty with my current software. However the biggest disadvantage is the need to still push a switch to engage drive. There's a good reason for this but it makes it impossible for me or others with no purposeful movement left, to use. So Patrick was determined to develop a purely eyegaze controlled system. I was sceptical about the idea knowing some of the limitations of eyegaze. It was for that reason Patrick invited me to assist developing the system. He concluded if I could be convinced then anyone could.   

The system comprises a mechanical hand which is moved by tiny servos. This was manufactured by Patrick using his 3D printer. Then there's a processing unit, again constructed and programmed by Patrick. The system connects to the eyegaze computer by USB. Finally a screen grid was made to allow eye selection. 

The system is still in its early stages of development but early tests are promising.

This video was of the first day of testing and shows the mechanical hand in action.




The next video was made in the seafront carpark at my flat




The two disadvantages I expected with eyegaze control were a delay in sending commands and the problem of sunlight. I select items by looking at the screen cell for a set "dwell" time. My dwell time is set to 0.7 seconds. This means there's a 0.7 second delay between wanting to start or stop and the wheelchair starting and stopping. During initial testing this delay was compounded by a 1 second processing delay. Almost 2 seconds is a long delay when you're heading towards a river or even the edge of a kerb. Anyway Patrick has eradicated the processing delay and as a result I am feeling positive about the next testing session later this week.






Ultimately eyegaze control will never be as responsive or accurate as hand control (as Patrick discovered last week at Centreparcs when his wheelchair made friends with a ditch) but I think in a safe environment it will be fine.... as long as the sun doesn't come out. But joking aside, just to able to turn or move the wheelchair indoors will be a HUGE improvement over our current static situation. Ultimately we want to make this system available to anybody for just a couple of hundred quid.