MIT-algorithm-300x200

Image: MIT News

Researchers from the Computer Science and Artificial Intelligence Laboratory (CSAIL) at Massachusetts Institute of Technology (MIT) have presented a new lane-changing algorithm for autonomous cars. This creates and reassesses ‘buffer zones’ around a vehicle and is said to allow for more aggressive lane changes than existing simple lane-change algorithms, which tend to be impractically cautious and conservative, yet is not too complex for on-the-fly analyses.

The algorithm relies only on immediate information on other vehicles’ directions and velocities to make decisions. “The motivation is, ‘what can we do with as little information as possible?’ How can we have an autonomous vehicle behave as a human driver might behave? What is the minimum amount of information the car needs to elicit that human-like behaviour?” explains researcher Alyssa Pierson, lead author of a research paper presented at the International Conference on Robotics and Automation.

It goes beyond current pre-computed buffer zone calculations - which describe surrounding vehicles’ current positions and also their likely future positions – with its more dynamic on-the-fly evaluation using a simple mathematical model created by the team. This was tested in a simulation including up to 16 autonomous cars driving in an environment with several hundred other vehicles.

Pierson concludes: “The autonomous vehicles were not in direct communication but ran the proposed algorithm in parallel without conflict or collisions. Each car used a different risk threshold that produced a different driving style, allowing us to create conservative and aggressive drivers. Using the static pre-computed buffer zones would only allow for conservative driving, whereas our dynamic algorithm allows for a broader range of driving styles.”

The project was part-supported by the Toyota Research Institute and the Office of Naval Research. The research paper, “Navigating Congested Environments with Risk Level Sets” (Pierson, Schwarting, Karaman and Rus), is published in the Proceedings of the IEEE International Conference on Robotics and Automation (ICRA), May 2018.

-Farah Alkhalisi