For this week, I worked on multiple aspects of our project. First, I validated the rotational data using the glove. In order to perform the test, I developed a way of using Processing to save all the rotational data generated with serial. This is a crucial step for testing because I collect rotational data every 5 milliseconds and there is no way for me to keep track of all the data. So the above implementation helps all those serial outputs to be saved in a text file. Afterward, I used Python to iterate through a text file line by line to find the rotational data that I want to use to validate my testing. The test procedure was to rotate my hand with the glove on for 100 trials. Then, I cleaned up the serial outputs to only the last rotational data per instance. By doing so, I was able to find the angles the sensor produced every time a user thought that they made a 90-degrees hand turn. Using 100 output values, I calculated the standard deviation with the population-mean equal to 90. The test resulted in a standard deviation of 5.74 which meant that for every 100 trials, all the output angles were within the range of +- 5.74 of 90-degrees. This measure gave us confidence in how stable the rotational data is and we noticed the great performance when seeing the rotational data changed the parameter values in the software synthesizer.
Second, I worked on designing the tray for the Jetson, Due, and MIDI processing board circuit. Tomas and I visited TechSpark to talk with one of the faculty members there to learn about the best method of making the tray. Justin, the faculty member, suggested that we laser cut an acrylic plate to make the tray. However, we found out that there are some safety measures to learn before using the laser cutting machine. So, Tomas took the online course regarding the safety measures and I made the CAD file. We are meeting again with Justin tomorrow to use the laser cutter and finalize the tray.
Third, our team did two user validation tests yesterday. Upon the tests, we came up with a structured procedure to perform user validation tests. We made a Google form so that users can fill out the form after their tests and also came up with questions to ask during the test sessions. Two users who came in yesterday were electronic music students. They really enjoyed using our product and they also gave us some suggestions for improvement. We will be doing another test tomorrow and then the team will look into what improvements can be made before the demo day.
As of today, my schedule is on track and our team is keeping track of all the work that needs to be done before the demo. I still have some leftover work to do to refine the velocity data of the glove. Tomas suggested a method of making the current complementary filter to be in a higher order. This is the approach that I will be looking into and hopefully have it working by next Monday. Also, I will come up with ways to finalize the assembly of the glove next week.