Speed:
0 mph
Cars Passed:
0

Road Overlay:

Simulation Speed:

DeepTraffic

Visualization - Leaderboard - Documentation - Paper - GitHub

Americans spend 8 billion hours stuck in traffic every year.
Deep neural networks can help!



DeepTraffic is a deep reinforcement learning competition. The goal is to create a neural network to drive a vehicle (or multiple vehicles) as fast as possible through dense traffic. What you see above is all you need to succeed in this competition. Here are the basic steps:

  1. Change parameters in the code box (read documentation for hints).
  2. Click "Apply Code" white button.
  3. Click "Run Training" blue button.
  4. Click "Submit Model to Competition".

In version 2.0, you can create visualizations of your best submission:

  1. Customize your image vehicle (Load Custom Image).
  2. Customize your color scheme (drop down).
  3. Click "Request Visualization".

The result will look something like this:

Some other DeepTraffic pages you should look at next: Visualization, Leaderboard, Documentation, Paper, GitHub. Or connect with Lex (@lexfridman) on Twitter, LinkedIn, Instagram, Facebook, and subscribe on YouTube.