One of the most significant risks to our project is ensuring that we have enough time to train a machine learning model on sensor input data. If we do not set up our data collection system early, we may not have sufficient training data, which could result in an underdeveloped model that does not perform as expected. To mitigate this risk, our approach is to prioritize setting up the fundamental hardware and software components as soon as possible. This includes setting up the Raspberry Pi 5 with an initial sensor and the camera module, establishing a data collection and storage system to begin logging sensor and camera input, and implementing the initial machine learning pipeline so we can start training models early in the development process. By doing this, we ensure that data collection and ML training can happen concurrently with other parts of the project, minimizing delays.
There were no major changes to the design of the system. However, we made one key adjustment by switching from the Raspberry Pi 4B to the Raspberry Pi 5. This decision was made because the Raspberry Pi 5 offers better performance, which will help with machine learning training and real-time processing. Additionally, we were able to secure one for free from the ECE inventory, reducing project costs. This change does not introduce additional expenses but improves our system’s capability and allows us to work with newer hardware.
At this point, there are no major schedule updates. Our current focus is on ordering components and beginning initial setup. Once our parts arrive, we will start hardware integration and software development.
So far, we have finalized our project proposal and user case requirements, researched communication protocols for sensor and camera data transfer, ordered our first Raspberry Pi 5 unit, and finalized the initial list of parts to order next week. Our next steps include ordering all remaining parts and beginning hardware setup, setting up the Raspberry Pi 5 with essential sensors and the camera, and establishing the data collection and machine learning training pipeline. We will include photos of our hardware setup in future updates as we begin assembling components.