MIT Self-Driving Cars Can Drive on All Roads with No 3D Maps

This, however, doesn’t stop researchers from other organizations from trying to find better solutions to the problems this type of vehicles faces.

One of the biggest challenges for AVs is finding the correct path down a road. For this, they use complicated 3D maps of the surroundings, preferably live-updated so that they can stay on top of any situation. But, as researchers from MIT say, the roads included in these maps are only a tiny fraction of the existing network.

So MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) developed a solution that would allow AVs to navigate their way without using 3D maps. It is called MapLite and combines simple GPS data with a series of sensors that observe the road conditions.

Tests of the system have been made on multiple unpaved country roads in Devens, Massachusetts. The car fitted with MapLite – a Toyota Prius equipped with LIDAR and IMU sensors - was capable, according to MIT, to detect the road more than 100 feet in advance.

“The reason this kind of ‘map-less’ approach hasn’t been done before is because it is generally much harder to reach the same accuracy and reliability as with detailed maps,” said Teddy Ort, CSAIL graduate student.

“A system like this that can navigate just with onboard sensors shows the potential of self-driving cars being able to actually handle roads beyond the small number that tech companies have mapped.”

The full paper, which you can find in the document attached below this piece, will be presented this month at the International Conference on Robotics and Automation (ICRA) in Brisbane, Australia.