DARPA’s RACER program has been developing off-road autonomous combat vehicles that can travel as fast as crewed systems. The agency has now selected Intel to develop a simulation platform for these off-road AVs. The aim of the project, known as RACER-Sim, is to create computer models that mimic the type of rugged, unstructured terrain that these vehicles commonly encounter. Intel Labs will work with the Computer Vision Center in Barcelona and the University of Texas at Austin to craft these simulation tools.
“We brought together a team of renowned experts from the Computer Vision Center and UT Austin with the goal of creating a versatile and open platform to accelerate progress in off-road ground robots for all types of environments and conditions,” said Intel’s Autonomous Agents Lab director German Ros in a statement.
Developing AVs that can operate on backroads, hilly areas and other types of rough terrain have been a major challenge for the industry. Traditional lidar and camera sensors are designed for predictable settings like roads and highways. Off-the-road AVs must operate in a much more challenging environment, driving on extreme terrain with rocks, debris and vegetation.
Over the next two years, Intel will work on speeding up the process of designing off-the-road combat AVs. During the first phase, it will create new simulation platforms and map generation tools that can mimic off-the-road environments. Intel during the second phase will implement the new algorithms without relying on robots, a measure designed to cut costs and time.
The Pentagon for many years has been working on autonomous technologies for the U.S. military, including unmanned underwater vessels, pilotless fighter jets and more. But the cybersecurity and safety risks of such systems pose a real challenge to their everyday use.
You can watch a preview of what RACER-Sim will look like in the video below: