Team status report for 4/4/26

This week we mainly worked on the demo portions of everything. This meant that we fixed some small portions of our project. Initially, we could send a request to upload an image, but our metadata was not syncing. We spent a couple hours debugging and eventually found the root cause. For the CV, we spent time retraining the model on some better data to make sure that we were more accurate. We also connected the external GPS module to our RPi, so now our whole product is connected. We are now starting to test the different parts of our project. This means that we are doing things like latency testing, unit testing, and other user testing. We wanted to work on connecting our Pi to the car, but due to some personal reasons, we won’t have access to the car for a few days. This won’t be an issue because we have everything ready, we just need to plug it in. We hope to be done with the product, minus the testing this week.

 

For verification, we have worked as individuals to make sure our individual portions work, such as benchmarking our CV with our test data or load testing our backend. We also verified that our product can run without being plugged into an outlet, but through our actual UPS. For validation, we want to run multiple user tests with different cars. This way we can ensure that our product is easy to install and simple to understand what is happening. We also want to see if people agree with what we label are “severe” potholes.

Kevin’s Status Report for 4/4/2026

This week, I started testing our web app with different data points. I also helped with integrating the GSP transponder onto our hardware RPi. We also started to test our hardware in real life to see how our hardware performs under different conditions. In the next week, I hope to finalize the entire software flow after testing and then continue to work to finalize the hardware and then field/user testing. This is on schedule with our proposal.

In terms of testing, I worked on the software web app portion of our application and how I envisioned to test this was to input a lot of dummy data and see how my web app manages scale. Further on, once our full pipeline is robust enough I will want to see that the data extracted matches the field data we have gathered. This data can also be cross referenced with manual in-real-life comparison.