CollisionSpect

Collision Scenario
Crash Prediction and conflict
Home Screen

Inspiration

Traffic lights are one of the most dangerous parts of driving. According to the AAA, in 2017, 939 people were killed in red light running crashes. Red light cameras don’t solve the problem because they are reactive and not proactive. Many red light runners don’t even realize they’ve run a red light, and once a collision happens, it doesn’t matter if there was a camera or not.

So, we thought, if Google maps can give data on bad traffic so well, there’s no reason we can’t use similar data to prevent accidents. Taking inspiration from the Traffic Collision Avoidance System (TCAS) on airplanes, we wanted to use the same concept of predicting the course of the two vehicles and giving instructions to each if they’re on a collision course.

What it does

Our idea uses the live location, bearing, and speed, as well as the change of such data over time, for the current vehicle and all nearby vehicles, to predict if a car will run a red light and if that car has a chance of colliding with you. If this occurs, both drivers will be given an alert to stop, such that at least one of them is able to, avoiding the collision.

How we built it

We used HyperTrack’s API to get the user's geoJSON location which includes all the information we need. We also used Flutter and Dart to build the mobile user interface.

Challenges we ran into

Testing out our app was challenging since the live location data within the building was not accurate. There is also a slight delay for the GPS coordinates to be updated and for the warning to be alerted. Working with cross-platform mobile development was also challenging.