Update: Working with Glass

Sorry it's been so long since I updated this blog. I finally got my hands on Google Glass just over two weeks ago. In that time things have moved a lot quicker now that I can actually test out some code. Google is working on a Glass Development Kit, but in the meantime has encouraged developers who can not rely on the Mirror API to use the Android Development Kit to write their programs. Any applications written with the Mirror API currently cannot access the device hardware, such as the Bluetooth receiver, and rely on hitting a server for each action, which would be dangerous for wheelchair control.  The following images are all screenshots from my Google Glass using Det Ansinn's launcher.

The first program I wrote for Glass was basic application to test moving between different actions in the Glass UI. I'm starting to play around with voice recognition and the acelerometer as two separate hands-free input mechanisms to make up for the lack of eye-facing camera. The place that would have the camera, according to Google's patent (US 8,235,529), instead has a light sensor, but Catwig's teardown makes it seem as though that could be replaced by a small camera in a future release. For the time being, I'll just abstract out the input so that the user can pick which one is best for them and plug in eye-tracking as another option when it becomes available.

The other program I wrote was a driver for listening to the Bluetooth receiver for input and sending out basic commands over Bluetooth. I installed this same program on both an old Android phone and Glass and was able to get them to speak with each other. The next step will be to get the small Arduino car I set up to receive the Bluetooth signals and move its motors based on the received messages. This should take a bit of time given the added complexity of getting Bluetooth setup on Adruino, but it'll a ton of fun. It's also a major step towards getting this up and running. Once I can translate messages sent from Glass into motor movements, the next steps are refining the voice and accelerometer input mechanisms and coming up with a polished UI experience for the user.

I'll post soon about the progress I'm making towards getting the car moving around.