MIT project helps self-driving cars navigate rural roads

How MIT's car sees the world with a LiDar sensor (MIT illustration)

Self-driving cars are already on the streets in a few urban areas, but they face a significant barrier to driving rural roads: there may not be detailed three-dimensional maps available (think Google Street View maps), and the roads can be twisty with few or no buildings nearby to help the car assess landmarks or road edges. But researchers at the Massachusetts Institute of Technology have come up with a way to help autonomous cars navigate rural roads, and "their strategy involves teaching cars to drive like humans," Rob Verger reports for Popular Science.

Though self-driving cars rely on detailed maps to help them figure out where they are in a city, many have a LiDAR (Light Detection and Ranging) sensor that helps it see curbs and other obstacles. The MIT team juiced up their test vehicle's LiDar sensor to help it detect the difference between the asphalt and the grass. The LiDAR's 64 lasers each spun around 10 times per second, providing constant feedback on what the car's surroundings look like and where it was on the road.

That’s how the car perceived where the road was in front of it, but it still had to know how to drive to its destination without a great 3D map in its silicon brain (although it did have GPS)," Verger reports. "To do that, it picked a 'local goal'—a point in the road up ahead that the car could see, and drove towards it. But it didn’t just drive to that point and stop. The vehicle constantly refreshed that goal as it approached it, like paddling towards a point on the horizon on a big, flat lake."

Christoph Mertz, a scientist at Carnegie Mellon University's Robotics Institute, admires the project because he says rural areas are sometimes neglected by scientists, and this project could help rural residents: "If these autonomous vehicles don’t drive in rural areas, then the elderly there might be stuck in their houses because nobody can drive them."