Self-driving technology is still in its infancy. And one main reason why self-driving cars are tested on cities, is because they need well-labeled 3D maps to identify things like curbs, lanes and signs.
"The cars use these maps to know where they are and what to do in the presence of new obstacles like pedestrians and other cars," said Daniela Rus, director of MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL). "The need for dense 3D maps limits the places where self-driving cars can operate."
Here, CSAIL with the help from Toyota, has developed MapLite. The goal of this system, is to create autonomous cars that can navigate on any road without anyone holding the steering wheel. To do this, the framework enables self-driving cars to navigate its ways through roads that have no 3D maps available.
The system uses Google Maps and GPS to know the location of the car, and use its position on both the final destination and the "local" objective within the view of the car. The system then use other sensors has car has to generate the path to those local points.
The car can then use its LiDAR and pulse laser lights to identify distances and estimate the edges of the road.
Country roads can cause problems for autonomous cars for a number of reasons. Most notably is because the roads are poorly marked and also difficult to map. And the reasons why there is little incentive for people to map them, is because few people ever want to drive their cars on them.
Companies like Google only test their car fleets on places where people go often. They also update their maps by spending a lot of resources and times to label the exact 3D positions of lanes, curbs, off-ramps, and stop signs.
"The reason this kind of ‘map-less’ approach hasn’t really been done before is because it is generally much harder to reach the same accuracy and reliability as with detailed maps," said CSAIL graduate student Teddy Ort, who was a lead author on a related paper about the system. "A system like this that can navigate just with on-board sensors shows the potential of self-driving cars being able to actually handle roads beyond the small number that tech companies have mapped."
With MapLite, CSAIL wants to change that. MapLite allows self-driving cars to adapt on the fly. As a start, the car can reliably detect the road more than 100 feet in advance.
"I imagine that the self-driving cars of the future will always make some use of 3D maps in urban areas," continued Ort. "But when called upon to take a trip off the beaten path, these vehicles will need to be as good as humans at driving on unfamiliar roads they have never seen before. We hope our work is a step in that direction."
In its initial state, MapLite can't handle everything by itself. For example, it doesn't know how to cope with sharp twisting roads on mountains, or know how to adapt the car on sharp changes in elevations.
