Inspiration

Every day, time was wasted on roads at intersections because of the unreasonable distribution of duration of green lights. If the roads with more cars can have longer durations of green lights, people's time can be saved.

What it does

Our model dynamically changes the duration of green lights according to some features (like the volumes of cars on road, the expected upcoming cars per second and so on) to minimize the total wait time at the scale of a city.

How we built it

We use Amazon AWS as our cloud server, building our model using CNN(Convolutional neural network).

Challenges we ran into

We ran into a few challenges while developing this project: The data-sets which we utilized were very dense and overwhelming at first but as a team, we were able to attack them together and succeed in analyzing the data. Another challenged we faced involved some additional factors that needed to be taken into account in order to effectively complete this project such as Pedestrians, cyclist, left/right turns etc. Although these were challenges, they did serve as a sort of acceleration toward our main goal

Accomplishments that we're proud of

What we learned

In order to address the larger problem, we can set some limitations first and try to solve the simplified problems first.

What's next for Traffic Congestion with Smarter Signal Lights

• Scale the simple problem solution on a whole city. • Take into consideration pedestrians, cyclists, STOP signs etc. • Extend the model for Left/right/U turns • When: Weekends, rush hour etc. • Update frequency of real time scaling. • Provision for Emergency vehicles: prioritized and non-prioritized