Wheelchairs can be pushed. Wheelchairs can be controlled by joysticks. But what about a wheelchair for someone who is paralyzed from the neck down? With approximately 66 million people paralyzed worldwide, it is important to grant the power to control movement to those who have extremely limited motor functioning.
What it does
Our wheelchair is powered through the simple facial motion of raising one's eyebrow, a skill that can be done even if someone is completely paralyzed from the neck down.
How we built it
We used the Xilinx PYNQ board to integrate python code, a usb camera, and output to our wheelchair. The camera first takes a baseline image of the user's face to calibrate. After that, it continuously takes pictures every half second to detect raised eyebrows. Pictures are processed with the FacePlusPlus machine learning classifier in order to detect facial landmarks. Calculations on the landmarks returned from the classifier enable the board to detect raised eyebrows. Passing the image from the camera to the classifier involves an ftp protocol and an HTTP Post request to the FacePlusPlus API. We also used a preexisting wheelchair base to build up our chair. After fixing the motor and fashioning a new chair, we were able to wire the PYNQ and arduino to the motor of the chair.