Rapid Autonomous Complex-Environment Competing Ackermann-Steering Robot (RACECAR) 2015


Lincoln Laboratory Instructors:

MIT Faculty Advisors:

Michael Park, Michael Boulet, Owen Guldner

Sertac Karaman

RACECAR Results:

  • Three of the four teams were able to complete the course
  • The fastest time was 49 seconds (average speed: 7.0 mph; Team 4)
  • 2nd Place was 57 seconds (average speed: 6.0 mph; Team 1); 3rd place was 118 seconds (average speed: 2.9 mph; Team 3)
  • Predominant navigation strategy: Use 2D LIDAR data to detect obstacles, corners and open spaces
    • Bias of open spaces given to resolve decision between which corridor to choose when one is found (e.g. choose the left path for a counter-clockwise closed loop)

Goals for the Participants:

  • Have hands-on experience developing dynamic autonomous systems, including familiarity with sensing, embedded processing, and autonomy software
  • Have knowledge of autonomous systems challenges and solution approaches, and in particular, issues associated with increasing the speed of robots in complex environments
  • Be able to use the ROS (Robot Operating System) API to develop advanced robotics software on a diverse array of unmanned and robotic systems 
RACECAR Course in MIT Strata Center

Class Structure:

  • 5-6 students that composed four teams
    • Teams consisted of a mix of undergrad and grad students in MechE, Aero/Astro, CSAIL and SUTD
    • Prerequisite for IAP course: intermediate level of programming and control algorithms experience
  • Lectures and Hackathons
    • 12 session in January 2015 at BeaverWorks, Aero/Astro Hanger and Strata Center
  • MIT Tunnels below the Strata Center
  • Team that completed the course in the shortest amount of time was the Winner


  • The class will be offered as a Technology Education course within Lincoln Lab in Fall 2015
  • Potential for making RACECAR an extension of the previously offered Rapid Robotics course
    • Things to improve for the next iteration:
    • Stronger mechanical mounting of the electronics to the chassis
    • Work to tailor the lectures to be more in tune with the technical content of the course
    • Implement a SLAM-based and/or laser scan matching solution