MIT built a self-driving car that can navigate unmapped country roads

Taking the road less traveled is extremely difficult for driverless cars. Autonomous vehicles rely on highly visible lane markings, as well as on detailed 3D maps to navigate safely in their environment. That is why most major companies have avoided testing on unmapped rural roads in favor of suburbs and cities.

Researchers at the Computer Science and Artificial Intelligence Laboratory (CSAIL) at MIT have developed a new system that allows car drivers to drive on roads they have never been to before without 3D maps. Called MapLite, the system combines simple GPS data that you would find in Google Maps with a series of sensors that observe road conditions.

This allowed the team to drive autonomously on multiple unpaved rural roads in Devens, Massachusetts, and reliably detect the road with more than 100 feet of anticipation. (As part of a collaboration with the Toyota Research Institute, the researchers used a Toyota Prius that they equipped with a range of LIDAR and IMU sensors.)

"Cars use these maps to know where they are and what to do in the presence of new obstacles such as pedestrians and other cars, "says Daniela Rus, director of CSAIL at MIT, in a statement. "The need for dense 3D maps limits the places where autonomous cars can operate."

MIT explains how your system works without relying on three-dimensional maps:

MapLite uses sensors for all aspects of navigation, based on GPS data only to obtain a rough estimate of the car's location. The system first establishes both a final destination and what researchers call a "local navigation objective," which must be in view of the automobile. Their perception sensors then generate a path to reach that point, using LIDAR to estimate the location of the edges of the road. MapLite can do this without physical markings on the road by making basic assumptions about how the road will be relatively flatter than the surrounding areas.

If it ends up being commercial, MapLite from MIT would do much to fulfill one of Trump's administration mandates: that the security benefits of autonomous driving extend to residents of rural communities.

In Detroit last year, the secretary of EE. UU., Elaine Chao, said it is vital for the auto industry to ensure that cars that drive autonomously help improve life in rural communities and not just in urban areas. "We want to be inclusive and consider how this technology can benefit rural America," said Chao during the 2017 Detroit Auto Show.

The question is: what is the use case for driverless cars in rural communities? Most of the leading companies, such as Waymo and GM, pursue commercial businesses that use autonomous vehicles. Robotaxis is considered the best way to bring autonomous technology to most people in the shortest possible time.

It is also likely to be the application that generates the most money for manufacturers. And the economy of horse travel really does not work in rural communities. Like Uber or Lyft, which are fairly non-existent in those communities. If it is shown that the cars that drive by themselves are of a much safer magnitude than those driven by people, then we are likely to see a stronger push to expand their capabilities to all communities.

MIT presented its research findings in May at the International Conference on Robotics and Automation in Brisbane, Australia. The document was co-written by CSAIL director Daniela Rus, graduate students Teddy Ort and Liam Paull.