This video illustrates some of the types of terrain that our autonomy stack must navigate through. This section of the course includes rocks hidden amongst bushes, large rock piles, joshua trees, and interesting topography. In the middle of this run, the vehicle opportunistically finds a trail and takes advantage of it to increase speed.
The objective is to move as quickly as possible between checkpoints, which are marked by flags. The autonomy stack controlling the vehicle does not use any advance maps or other prior information about the terrain; all navigation is done using on board sensors and compute.
Since November 2021, University of Washington researchers from the Paul G. Allen School of Computer Science and Engineering and the Applied Physics Laboratory have been advancing autonomous driving in off-road terrain with novel approaches to perception, planning, and control. UW has been developing this work under DARPA’s Robotic Autonomy in Complex Environments with Resiliency (RACER) program.
