Inspiration
A law was passed that made it illegal in Texas to use traffic cameras to enforce traffic laws so it would be good to repurpose them to alleviate traffic congestion. This could save emergency vehicles seconds or even possibly minutes to get where they are needed most.
What it does
It combines a neural network with computer vision to manage the control system of an intersection's traffic lights to reduce congestion.
How we built it
We built the main functionality using Python but we also used C#/Unity to visualize what our project does.
Challenges we ran into
Training supervised model under these time constraints and finding a good model for vehicle detection.
Accomplishments that we are proud of
We achieved significant improvement in our simulation and we show how much room for improvement there is in the real world.
What we learned
How bad traffic congestion really is and how to work with supervised models under a time constraint.
What's next for Light Savers
Possibly pitch the idea to cities!
Log in or sign up for Devpost to join the conversation.