Project
| # | Title | Team Members | TA | Documents | Sponsor |
|---|---|---|---|---|---|
| 55 | Head-Motion Controlled Wheel Chair |
Arnav Jain Dev Manaktala Jiayuan Liu |
William Zhang | design_document1.pdf design_document2.pdf other1.pdf proposal1.pdf |
|
| #Problem Statement Classic electric wheelchairs are controlled by a knob at the hand. They require people to use fingers to operate them. People who are unable to move their hands or arms (who suffer from paralysis or are amputees) cannot benefit from this type of wheelchairs. A new type of wheelchair becomes necessary to serve for people who can move their head but are unable to move their hands or arms. #Solution Overview We propose an electric wheel chair with a mounted camera that captures the movement of the user’s head. It would use computer vision to detect the motion of the user’s head and give feedback to the motor controller accordingly. We plan on using a raspberry pi to interface with the motor controller. We would ideally like to use edge computing to process the camera footage on a GPU to accelerate the process. Instead of placing motors on the pre-existing wheels of a regular wheelchair, we would add two more wheels at the back that can move and turn the wheelchair. The PCB will integrate the raspberry pi and two sets of motor control circuits to achieve both moving straight and turning by adjusting the speeds of two motors according to the signals received from the raspberry pi. Additionally, we plan on adding ultrasonic/ IR sensors to detect objects in front of the wheelchair and provide vibration feedback using a vibration motor. All features will be implemented by a single PCB. Our solution would solve the problem addressed as it would give people suffering from paralysis or amputees the ability to transport themselves. We differentiate ourselves from products currently available by using head motion and adding sensors for object detection which would give people a better understanding of their surroundings and add another layer of safety. #Solution Components Components Required: Two motors, one camera module, two wheels, one wheelchair, Ultrasonic/ IR sensors, vibration motors, a Raspberry Pi, and a 40V battery pack. -Motion Detection Subsystem: Footage from a mounted camera would be sent to a raspberry pi where the head motion detection would take place. We will be running Haar Cascade facial detection and then track the movement of the bounding box as the person tilts their head. This would then send feedback to the motor controller. -Motor Controller Subsystem: Two motor control logics on the PCB receives signals from the motion detection subsystem and adjust the speeds of two motors to achieve both moving straight and turning. The motor controller will be powered by a battery pack at around 40V. -Wheelchair: A wheelchair will be modified to accommodate the electronic system, and more importantly, the battery pack and two motors. -Object Detection subsystem: Sensors would measure distance to objects near the wheelchair and send that data to a microcontroller. This would process data and accordingly send signals to the vibration motor. #Criterion for Success Our solution will be successful if the wheelchair is able to go forward, stop, turn left, and turn right according to four motions of the user’s head. Users are also able to get vibration feedback when surrounding objects are detected. #Partners Arnav Jain Dev Manaktala Jiayuan Liu |
|||||