Science

How MIT’s MapLite Will Bring Autonomous Cars to Donald Trump Heartlands

Autonomous cars are going mapless. The Massachusetts Institute of Technology’s Computer Science and Artificial Intelligence Laboratory revealed MapLite on Monday, a groundbreaking approach to self-driving vehicles that allows systems to drive on roads without the aid of an existing 3D map. The feature could enable the likes of Waymo and Uber to offer a self-driving taxi service that expands beyond the limits of a city — breaking past the “level four” systems expected to hit roads by 2021 that can drive around a pre-defined area, to reach the holy grail of a “level five” car that can drive itself anywhere in any conditions.

“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,” says CSAIL graduate student Teddy Ort, a lead author on a related paper. “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.”

MapLite provides the missing puzzle piece to enable widespread autonomy, crucial for the millions of miles of American roads with no markings, paving or lighting. Companies have so far limited autonomy to a small test area, like Waymo’s Phoenix project or Uber’s Pittsburgh trials. MapLite could bring the technology to the rust belt heartlands of Donald Trump’s election victory, rural regions that typically benefit less from these city-focused breakthroughs in transport.

The group's test car.

MIT

The system uses GPS to roughly estimate the car’s position, then uses lidar to estimate the car’s surroundings. The user sets a final destination, and the system regularly chooses local navigation goals in view of the car. MapLite judges the edges of the roads by making a few assumptions, like the fact that the road is probably flatter than the rest of the area. A series of models enable MapLite to judge how to approach new situations, like approaching an intersection.

The team consists of Ort, professor Daniela Rus and PhD graduate Liam Paull. The paper, entitled “Autonomous Vehicle Navigation in Rural Environments without Detailed Prior Maps,” will be presented at the International Conference on Robotics and Automation in Brisbane, Australia, this month. The National Science Foundation and the Toyota Research Initiative both supported the project.

“Our minimalist approach to mapping enables autonomous driving on country roads using local appearance and semantic features such as the presence of a parking spot or a side road,” says Rus.

The team still has a lot of work to do before MapLite can reach the masses. It struggles with mountain roads, as it cannot process the large changes in elevation. The team wants to develop the system further so that it can reach a level of performance comparable to mapped systems. While they don’t envision a future vehicle that ditches the maps for good, MapLite could complement a map-based system to enable trips further afield when required.

Read the paper abstract below:

State-of-the-art autonomous driving systems rely heavily on detailed and highly accurate prior maps. However, outside of small urban areas, it is very challenging to build, store, and transmit detailed maps since the spatial scales are so large. Furthermore, maintaining detailed maps of large rural areas can be impracticable due to the rapid rate at which these environments can change. This is a significant limitation for the widespread applicability of autonomous driving technology, which has the potential for an incredibly positive societal impact. In this paper, we address the problem of autonomous navigation in rural environments through a novel mapless driving framework that combines sparse topological maps for global navigation with a sensor-based perception system for local navigation. First, a local navigation goal within the sensor view of the vehicle is chosen as a waypoint leading towards the global goal. Next, the local perception system generates a feasible trajectory in the vehicle frame to reach the waypoint while abiding by the rules of the road for the segment being traversed. These trajectories are updated to remain in the local frame using the vehicle’s odometry and the associated uncertainty based on the least-squares residual and a recursive filtering approach, which allows the vehicle to navigate road networks reliably, and at high speed, without detailed prior maps. We demonstrate the performance of the system on a full-scale autonomous vehicle navigating in a challenging rural environment and benchmark the system on a large amount of collected data.
Related Tags