2017 CPS Challenge

The CAT Vehicle Challenge was a Spring 2017 student competition focused on model-based design for a self-driving car.

The competition aimed to broaden participation in Cyber-Physical Systems, and was run in the spirit of the DARPA Robotics Challenge as a competition in four phases. The Challenge provided an exciting opportunity for students to build a component that controls and runs on a real self-driving car—the University of Arizona’s CAT Vehicle.

Summary: In this challenge, participants were given data from the CAT Vehicle, driving along a path. The task was to use that data to identify potential obstacles along and off the path, and provide possible identification of those obstacles using as few sensors as possible. Each team selected to participate in the Final Challenge was be able to control the velocity of the CAT Vehicle as it drove along the trajectory. The result was a configuration file for a simulation environment (Gazebo) that mirrored the environment from which the data was extracted.

Tasks

The challenge had 4 Tasks:

• Task 1 (31 January 2017): Team registration with a submission to the CPS-VO that demonstrates competency to run the tutorials provided to all teams
• Task 2 (11 March 2017): Teams submit models and software to the CPS-VO that consume sensor and localization data to produce the coordinates of interesting objects after the conclusion of the CAT Vehicle's drive. Teams whose software executes successfully in simulation will have their code executed on the CAT Vehicle testbed, and will receive the resulting data files from their execution. Teams who are successful in Task 2 will be invited to submit to Task 3
• Task 3 (31 March 2017): Teams submit models and software to the CPS-VO that consume sensor and localization data to produce a desired velocity (while traveling along the trajectory), and within 60 seconds of stopping at the end of the trajectory, should generate a Gazebo world file centered at the origin point of the trajectory. Teams whose software executes successfully in simulation will have their code executed on the CAT Vehicle testbed, and will receive the resulting data files from their execution. The four (4) top scoring teams may be invited to participate in Task 4.
• Task 4 (Week of 24 April 2017): Teams will have an opportunity to modify and then re-run their Task 3 models on the CAT Vehicle in Tucson, AZ, over a period of 2-3 days. The models and software must be validated through the simulation framework before they are executed, for safety reasons.

Throughout each task phase, data and simulation traces and setups from the CAT Vehicle testbed was provided to teams through this group.