build details

Show: section status errors & todos local changes recent changes last change in-page changes feedback controls

Challenge LF

Modified 2020-11-07 by Andrea Censi

The first challenge of the AI Driving Olympics is “lane following” (LF).

In this challenge, we ask participants to submit code allowing the Duckiebot to drive on the right-hand side of the street within Duckietown without a specific goal point. Duckiebots will drive through the Duckietown and will be judged on how fast they drive, how well they follow the rules and how smooth or “comfortable” their driving is. Please refer to the following video of a lane following demo for a short demonstration. A description of the specific rules is provided.

A Duckiebot doing lane following

The challenge is designed in a way that should allow for a completely reactive algorithm design. This meant to say that to accomplish the challenge, it should not be strictly necessary to keep past observations in memory. In particular intersections will not be part of this challenge. Intersections will be recognized and maneuvered using provided code from the organizers.

LF in Simulation

Modified 2020-11-06 by Liam Paull

The current versions of the lane following simulation challenge are aido5-LF-sim-testing and aido5-LF-sim-validation. These two challenges are identical except for the output that you are allowed to see. In the case of testing you will be able to see performance of your agent (Figure 3.4) and you will be able to download the logs and artifacts.

Visual output for submission