Robocars Rule Book

Home / Robocars Rule Book

1. General Conditions

  • Teams up to 3 people.
  • Each team can have no more than one car of each class (see below).
  • We have a developer resources section on our website, with examples and source code that can be used to get started.
  • Each team will have equal access to the track for data collection and testing.
  • Any questions about the organisation of the challenge can be sent to robocars@deep-berlin.ai, please specify your team.

2. About the Model Cars (robocars)

There are two classes of robocars:

  •  Standard class: based on the ‘donkey carMagnet or HSP 94186  models (see specifications here), where participants must use the same chassis and camera. Only the computing unit can be extended or replaced and still remain under the standard class. If the sensors or the chassis are modified, the robocar will be no longer considered under the standard class but under the custom class (see below).
  • Custom class: there are no restrictions for this class. Participating teams that bring their own model fall under this class.

3. Robocars Provided by the Organisation

We have a limited number of standard class model cars available for talented teams that do not have one of its own. At the moment of filling in the challenge registration, the teams have to specify whether they need a robocar from us. Subsequently, the team will receive a second form that will assess the team members’ familiarity with certain programming and deep learning tools. Only then, the organisation team will make a decision after reviewing all applications, and announce the teams that will receive a robocar at the AI for Autonomous Mobility meetup happening on May 2nd, 2018.

The teams that apply for a model robocar and do not get the opportunity to receive one, will be contacted by us and receive the full list of components that they would need to assemble a donkey car and where to get them.

Important note: the models will be leased exclusively for the duration of the challenge, and therefore they will have to be returned to the organisation after the final race on September 12th, 2018.

4. Meetups & Testing

We organise and host the official AI for Autonomous Mobility meetup at the Hella Aglaia Mobile Vision GmbH offices once a month. We will provide the standard track configuration at every meetup for teams to perform test laps. The meetup events are an opportunity to exchange ideas and progress with other teams. For more information and updates on the event dates, times, and speakers join the meetup group here.

We will collect and make available the data from the track for participants who are not able to attend a meetup.

Thanks to Deep Berlin partners Fab Lab Berlin and MotionLab Berlin, the teams participating of the Robocars Challenge will have free access to their facilities during the period of the challenge, to work and experiment with their model vehicles.

5. The Final Race

There are two challenges:

  1. Lap race: the car has to complete 5 laps without leaving the track, the car that completes the fastest lap will be the winner.
  2. Lap race with obstacles: the winner has to complete a lap in the best time. Hitting an obstacle results in a time penalty.
  • Prizes will be announced.

6. Code of Conduct

As organisers, we would like to see a good spirit and motivation to not only build your own robocar, but also to be open to share ideas and engage in discussions with other teams. We expect a respectful attitude towards the spaces provided by Hella Aglaia and our partners Fab Lab Berlin and MotionLab Berlin.