7-in-7 | Day 4

For prototype 4, I explored my broader research question of enabling people with disabilities access spaces more easily and most of all, independently.

Some forms of muscular dystrophy can lose the ability to speak effectively (or at all). There are various devices that can allow them to speak, but most are extremely expensive and and usually are at least the size of tablets. This has to be mounted to a wheelchair and that creates another physical barrier with the world.

A common solution is called switch scanning (think Stephen Hawking’s screen mounted on his wheelchair, which means a grid of letters or words is displayed on a screen, and individual boxes are highlighted, one after another, and the user presses a button when the box they want is selected. Then the process repeats until the user is finished typing. My idea is that since a large screen is bulky and obtrusive, it would be better to use a small head mounted display instead. This presents a few challenges which I think I can solve. I plan to write a simple switch scanning program which will run on Raspberry Pi Zero which will be optimized for a miniature screen.

The main options I’ve explored are eye tracking and switch scanning. Eye tracking is a good option, but it can be difficult to position correctly, and means the screen has to be right in front of them. I have another idea for eye tracking, but each solution has its pros and cons. Switch scanning trades the difficulty of positioning for reduced speed.

I think using a small head mounted display or smart glasses would make a switch scanning system less obtrusive, easier to set up, and more portable. I’m trying a head/glasses mounted display called Vufine+, which is just a screen with no other features. It’s USD 200, which makes it a bit expensive, but compared to existing systems which can cost well over USD 5000, it’s much more affordable.

Challenges and Initial Solutions

• The display is supposed to be equivalent to a 4 inch screen at 1 foot away. So showing the entire grid at the same time would be unusable.
• The screen has to be mounted correctly to see the entire screen.
• The computer has to be very small, have a portable power source, and be able to output to speech and/or a screen. And a secondary goal would be for it to type to a computer via Bluetooth. Using a Raspberry Pi Zero W seems to be an optimal solution in this case.
• It must support a wide range of switches, because different users have different body parts that are strong enough. Most accessible switches use a 3.5 mm audio jack connector, so adding a 3.5mm jack to GPIO should do the job.