This week I verified all the components were working. I successfully set up the Jetson Nano, and got communications working with a Adafruit I2S microphone, multiple servos, and a Raspberry Pi V2 Camera Module. 

Setting up the camera module went smoothly and I can get a camera feed to the Jetson. When I get a second module this week, I think it will be an easy integration. I have also installed Jetpack to prepare for working on CV with Eddy this week.

The microphone set up was a little harder because documentation for I2S microphone to Jetson is scattered. After combing through the NVIDIA Jetson Linux Developer Guide and a couple of forums, I was able to set the correct commands for “amixer” and “arecord”. I2S codec is already installed on the Jetson and the GPIO pins have the option to be configured specifically for the Adafruit I2S Mem Microphone. I’m concerned that the Adafruit I2S microphone may not have a large enough range of audio pickup, but after some research, it seems like this could also be a part of the Jetson Nano software reducing the noise. 

Setting up the I2C motor control was fairly easy as well. I had some issues with the servo shield not being detected, but that was caused by some dumb wiring on my part. A plan of the design that we will have to change is our original idea for a continuous motor. After some research, a continuous servo can only be controlled by speed, however we need to have angle control. I am looking into a stepper motor, but I think we will most likely be switching our design to have servos that only have 360 degrees of rotation. I think this change will not greatly affect the overall project because we have multiple cameras. Something else to consider is currently, I have the motors connected to a 5V/2A power supply and the Jetson on a 5V/4A power supply. I think the power supplies need to be kept separate, as the current draw from the motors could cause the Jetson to shut down; however, this would mean our project requires two power outlets. I don’t think it is a pressing issue right now, but maybe later we can find a way to bring this down to just one outlet.

My goals for this week is to (1) get multiple cameras working, (2) make a final selection on motors, (3) work Eddy on the facial detection for CV, specifically with the goal of determining how the motor control integrates, (4) get a draft of the mechanical design, specifically the elevator part because I think this will be the most finnicky, and (5) work on the design presentation.