Team Status Report for 3/23/24

Potential Risks and Mitigation Strategies

We had some setbacks with the object detection model and wireless camera setup this week. For object detection, the original plan was to use Haar cascade classifiers to identify the number of traffic objects (cars, buses, pedestrians, etc.), but some of the functionalities we might need have been deprecated in the model we were using. In the case that this model is unusable for our purposes, we will instead look into using something like YOLOv4. As for the cameras, we needed to swap to a camera model that actually supports RTSP streaming, so we still need to receive the new camera and test that we can connect it properly to our RPi. If we have connection issues with the new camera, then we will default to using prerecorded videos as input to our system to demonstrate its functionality.

Changes to System Design

The only concrete change to the design this week is that we need to swap from the Reolink IP cameras to something that supports RTSP streaming. We are currently considering options from MubView and Amcrest and will have a replacement ordered this week.

Overall Takeaways and Progress

• We have made good progress on the simulation for optimization testing, and should be able to start gathering data with SUMO soon
• The PCB is ready to be ordered, and the other components for the traffic light circuit have already been ordered
• There are some unexpected challenges on the object detection front, but we have several backup plans in place and should still be able to demonstrate a system that meets our MVP requirements by the end of the course