We had our first zoom meeting on Monday (03/16)! We used the 10:30-12:30 lab time to begin to think about how we would transition our project to accommodate the remote nature of the class. We discussed which parts of the project we could keep the same and which parts would have to change, we also collected some questions we had for the course staff to help guide this decision. We also emailed professor Sullivan our availability and locations of team members. We began to think about how we could replace the hardware component of our project – potentially with an animation made through a platform such as SolidWorks? We were able to see how this would work pretty clearly for the manual mode of our project, but we were still working out the details of automatic mode. We briefly met with Jens to get checked off and ask him some questions about the new expectations for our project and he gave us a better picture of what we could do to prove that the software component of our project actually controls the hardware without requiring the hardware.

On Wednesday (03/18), we used the 10-30-12:30 lab time to further work out details for the transition. We spent the hour before meeting professor Sullivan coming up with a broad split-up of the work as well as questions we had for professor Sullivan about the course in general, the availability of parts, and the expectations for the Statement of Work document. We learnt that TechSpark would be open to take CAD designs and fabricate until the end of March and that all of our parts could be shipped to us. However, given the time this would take and the time we have lost already, we decided to stick with our plan of animating the output of our control loop. Shrutika and Gauri are both in Pittsburgh and once they are out of quarantine, they can meet off-campus to work on the hardware parts of our project. Neeti will be working on the classifier for the next couple of weeks as she is in New Jersey and does not have direct access to the parts.

On Thursday, Neeti did some research on existing gesture classification models and found the following links that will be helpful starting points:

https://towardsdatascience.com/tutorial-using-deep-learning-and-cnns-to-make-a-hand-gesture-recognition-model-371770b63a51

https://towardsdatascience.com/build-hand-gesture-recognition-from-scratch-using-neural-network-machine-learning-easy-and-fun-d7652dd105af

On Friday (03/20), we met from 10-12 to discuss the next steps Neeti will take in creating the classifier, we also did some further research into microphones and all the cardioid/unidirection mics were too large and/or expensive for our use case. After emailing professor Sullivan, We decided to use a baffle with a cheaper microphone. Finally, we outlined all of the decisions we have made in the SoW. We will come up with an updated Gantt chart and block diagram this weekend and submit the final SoW document.

We made it through the first week of quarantine! 🙂