As users of technologies with wireless charging capabilities, we wanted a seamless way for our phones to acutally charge wirelessly but always ran into issues when it came to actually charging our phone wireless.
Our objective was to develop a platform on which you could place your phone wherever you want on the platform, and the wireless charger would move to wherever the phone was placed.
- Stepper Driver Motor
- Gantry System
- Raspberry Pi
Essentially, we developed a software with the aid of OpenCV for a camera to detect the phone on the platform and then it would send the coordinates of the phone via Raspberry Pi to the Arduinos that we attached to each stepper driver motor (there were 2 of these in total: one for the x-direction and one for the y-direction). The Gantry system is the basic fundamental movement mechanism for our platform.
Some of the biggest challenges we face were actually developing an accurate enough software to detect the phone and not just detect any rectangles in general and to figure out how we send the transformed coordinates from image to real life to the Arduinos commanding the stepper driver motors. Additionally, we had to hard code the coordinate system for the system to use via trial and error which was a huge pain and just very tedious and time-consuming.
I think we came a long way since the beginning of the project and figured out how to overcome aforementioned challenges and actually come up with a prototype, although it is rudimentary and has many problems to fix and improvements to make. We're proud of building a project that we would use ourselves too, if not others facing the same problems as well.
What We Learned
We learned Arduino, OpenCV, and how to program these stepper drivers and setting the system up.
First, we need to find a way to prevent extensive light interference from background light with the camera that detects the phone in the first place. We can do this by rasterizing our platform and making it have a solid gray color which will make it easier for the camera to see the phone in the first place. Additionally, we need to make a way to point the camera looking down at the platform in the middle of it by adding standing rods on each of the 4 sides of the platform all of which support the camera looking down at a bird’s eye point of view. Furthermore, we need to improve our object detection software by possibly venturing into using deep learning methods and convolutional neural nets for feature detection and Haar Cascading to detect the phone better rather than just detecting anything that’s a rectangle because there are cases where there might be a rectangular shape in the field-view of the camera that isn’t a phone on the platform and our prototype won’t work.