The vast majority of American roadway miles haven’t been 3D-mapped in detail and/or can’t be reliably counted on to provide consistent lane markings. So a team from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) developed a way for autonomous cars to read and anticipate their local surroundings rather than relying on gigabytes of 3D map data.

The framework is called MapLite and it combines GPS using only the most basic topographic maps from OpenStreetMap with LIDAR and IMU sensors that monitor road conditions.