This week, the team’s major goals revolved around the design review. At the beginning of the week, we prepared our slides for the presentation, which Anna delivered on Wednesday. After that, we promptly began putting together our design review paper. Luckily, the solid effort we had put into the slides made the paper process much easier, as we already had a substantial template to work off of from the presentation.
A big takeaway from the design presentation was the realization that we didn’t have a full understanding of the setup for the communications between the Jetson Nano and a computer. Heather researched that this week and determined that we should be able to change the Linux kernel for the Jetson Nano using the Nvidia SDKManager. This change would be utilizing the Linux Gadget Driver and be made such that the Jetson Nano is recognized as a USB Video class, which is the same as a webcam.
We also had to make another change regarding the audio aspect of our design — last week, we had discovered that the four-microphone I2S setup would not work, since we need each individual microphone’s audio feed and thus cannot connect two to each of the Jetson’s I2S ports. Thus, we decided to use only two microphones, one on each I2S port. However, we soon realized that this presented a different problem: with only two microphones and calculating the difference in arrival time of the sound of the source to each, there will always be two possible locations of the source, on opposite sides of the microphone array. Therefore, we realized we would need a third microphone, which also means now having to use I2C (for which we have three remaining pins on the Jetson) instead of I2S, as well as analog microphones (and ADCs) instead of the digital I2S ones we were originally using.
In terms of the CV component, we were also able to make some headway in where we could use multiple cameras as input feeds for the computer vision so the next step is to start up on some more integration and the addition of implementing a software encoding zoom on the camera. The overall team goals for the week are to finish up the design report early on so we can continue working on the audio component and also the routing from the Nano to the computer.