Researchers at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) have successfully enabled a self-driving car to navigate over the remote country roads without 3D maps.
Daniela Rus, director of MIT’s CSAIL and his colleagues have developed a framework called 'MapLite' that allows autonomous cars to explore any roads without the availability of 3D maps. Maplite combines simple GPS data from Google with a series of sensors for observing the road conditions.
How does Maplite work
First, the handler of the car sets a destination that is visible to the car. The system relies on GPS data only to obtain a rough estimate of the car’s location. Then the perception sensors generate a path to reach to point using LIDAR to estimate the location of the road’s edges.
Maplite has been programmed to achieve this on those roads that do not have physical road markings. It can help the vehicle manouevre without the need of physical road markings and by assuming that the road is a much flatter surface as compared to its surrounding.
The MIT researchers collaborated with the Toyota Research Institute and National Science Foundation. The technology was retrofitted in a Toyota Prius with a range of LiDAR and IMU sensors.
Commenting on the project, Teddy Ort, a CSAIL graduate student and the lead author of the paper, said, “The reason this kind of ‘map-less’ approach hasn’t really been done before is because it is generally much harder to reach the same accuracy and reliability as with detailed maps. A system like this that can navigate just with onboard sensors, which shows the potential of self-driving cars being able to actually handle roads beyond the small number that the technology companies have mapped.”
According to the MIT team, currently Maplite isn’t completely reliable for autonomous driving on mountain roads, since it doesn’t account for dramatic changes in elevation. As a next step, the team hopes to expand the variety of roads that the vehicle can handle.
The paper, co-written by Ort, Rus and PhD graduate Liam Paull, will be presented at the International Conference on Robotics and Automation (ICRA) in Brisbane, Australia from May 21-25.