MIT researchers found a way to drive cars to the countryside by themselves

Cow? Cornfield? Self-driving car? "Country life" seems unrelated to high-tech traffic. But that may soon change, thanks to a new automatic navigation system developed by MIT researchers at CSAIL (CSAIL).

According to Futurism, today, most autonomous (AV) vehicles rely on detailed 3D maps for road navigation. The creation of these maps requires a serious commitment of time and money, so most AV producers are rushing to map the ever-clear roads, almost always shown shining. However, rural roads are not like that, not always clear, so these regions have never benefited from the technology for this future, even if the roads may be within communities need the most autonomous means.

The car is equipped with MapLite using a combination of GPS data and sensors to navigate the road.

Currently, CSAIL researchers have developed MapLite , a system that allows secure AV to navigate unmapped roads. They plan to present their project at the International Conference on Robots and Automation (ICRA) later this month.

Instead of relying on 3D maps, a vehicle equipped with MapLite uses a combination of GPS data and sensors to navigate the road. GPS provides the system with a basic idea of ​​the position of the vehicle, while the sensors allow "to see" the area around the vehicle, identifying the road edges based on the assumption that the road is flatter than the area. around.

Using the MapLite system, AV has "the ultimate navigation goal" (eg a grocery store), but also has a "local navigation goal". This second goal is always within the vision of the car's sensor, so it can be a point right in front of the road or an upcoming intersection.

The car will find the best way to reach your local goal, then set a new local goal.

As Teddy Ort, a graduate student CSAIL and the lead author of a related study, said in a press release, previous researchers had difficulty creating an AV navigation system. "little map" accurate and reliable.

Combining MapLite's GPS data and sensors is a unique approach to this problem. And seems to be effective.

When CSAIL team equipped MapLite for a Toyota Prius, AV successfully navigated many unpaved national routes in a sparsely populated town in Massachusetts.

"A system like this can be navigated only by sensors on the car, showing the potential of a self-driving car, can actually handle the roads that technology companies have not made on the map." , Ort said.

Finally, Ort thought that some combination of 3D and MapLite maps might be the best way for AV to navigate - 3D maps in urban areas and a MapLite-like system for unpublished roads. map. If so, the appearance of a self-driving car on a dusty dirt road may not be far away.