Inspiration
In our ever changing world, speed limits are one of the few constants. As a commuter, I find that to be a shame. Every day I pass speed limits that don't fit the current conditions, whether that's a baren, isolated road with a speed limit of 20 mph or a crowded street with pedestrians crossing with a speed limit of 55 mph. However, with the technology we designed in this hackathon, we could fix these issues.
What it does
Flow is a dynamic speed limit sign which self adjusts the speed limit based on various road conditions, such as amount of traffic and weather conditions.
How we built it
This project required a variety of Artificial Intelligence usage. One such case is the use of Yolo to detect the cars on the roads, allowing us to calculate the amount of traffic and check the speed of each car. We used Light GBM and turn it into a time series to predict future traffic flow. To obtain weather conditions, we used Meteomatic's API. The front end uses html, css, and javascript.
Challenges we ran into
A large challenge we faced was finding the appropriate footage to use to test our sign. While there's plenty of footage of cars driving, we couldn't find any at the height a speed limit sign would be at. Our solution was to traverse Nicolls Road to get the footage ourselves.
What we learned
We learned how to train our own AIs as well as dive into agentic coding.
What's next for Flow
Outside of giving it a physical form and not just being a theoretical program on a computer screen, we are looking to incorporate other factors into Flow, including visibility and car crashes. Additionally we would like to build a system where the police are notified whenever a car accident occurs.
Log in or sign up for Devpost to join the conversation.