This week, we connected all edges of the frame and spent most of our time debugging the hardware and software, as well as collecting data for measuring our performance against our use-case requirements. Beginning with the hardware, one issue we encountered was a misprint in the bottom PCBs. We were able to hack around this by soldering wires onto the board to fix the broken connection. Before we started testing our software with the frame, Matt S conducted some final debugging and checks to make sure there were no issues with the hardware.
Once the hardware was ok to go, we attached the frame to Matt S’ laptop and began testing of Matt K’s updated Arduino code, which controls the LED/Photodiode pairs and sends data about which pairs are activated to the python backend via USB serial. We encountered some bugs, but eventually after some squashing we were able to confirm that the control signals and output data were correct.
Afterwards, once we confirmed the Arduino subsystem was well-integrated with the hardware, we began our first true test of our entire system through a python script that used the data sent over by the Arduino to calculate coordinates and send commands to Darwin’s updated Touch Control subsystem. At first, a few bugs had came about due to a small miscommunication about the updated interface, but we were able to quickly fix them. Once we were passed that, we finally got our first glimpse at Touch TrackIR in action! Physical finger taps, holds, and drags on the screen were successfully translated to the expected touch inputs in the OS. We were able to click on links, close tabs, move windows, and scroll through webpages. The smoothness of the experience is acceptable, but we still feel there is room for improvement. For example, we can reduce latency by increasing baud rate, removing print statements, and by simplifying the structure of the data that is being used in our coordinate calculations to eliminate unnecessary loops.
Overall, thanks to the progress this week, we put ourselves perfectly on schedule. Next week, we plan on constructing a cover for our PCB frame and adding support for double taps to allow us to perform “right clicks”.