Mapless GPU method speeds drone obstacle avoidance in forests
Low-Latency Mapless ESDF Queries for MAV Reactive Collision Avoidance
Robotics
Summary
Flying drones safely through dense forests is hard because they must quickly detect and avoid obstacles like branches. The authors developed a fast method that uses camera depth images directly, without building detailed 3D maps, to figure out how close obstacles are. This approach runs efficiently on small onboard GPUs, making drones react faster and avoid crashes better than older methods that rely on pre-made maps. They tested this method in simulations and real forest flights, showing fewer collisions and higher speeds.
What this means in practice
- •For drone developers: Implement faster obstacle detection for drones flying in cluttered environments without prior maps, improving reactive collision avoidance onboard.
- •For robotics engineers: Integrate low-latency distance queries based on depth images into real-time robot motion planners for safer navigation through unknown terrain.
Authors
Leonard Freißmuth, Cedric Le Gentil, Helen Oleynikova, Stefan Leutenegger
Abstract
Reactive collision avoidance in unknown, cluttered environments must perceive obstacles and correct motion within milliseconds. Existing methods rely on maps for high-rate collision queries, but their acceleration structures require pre- processing. That cost is amortized in long-horizon planning, where a map serves millions of queries; reactive planning, however, is a near-field, short-horizon problem with potentially only hundreds of queries per measurement, where building a map may not pay off. We propose a mapless, Graphics Processing Unit (GPU)-accelerated Euclidean Signed Distance Field (ESDF) query with negligible pre-processing that operates directly on posed depth images, with resolution set by pooling the image rather than by a voxel size. On an embedded GPU it is simultaneously more accurate and lower in total latency than every voxel-grid configuration we measured between 2 cm and 20 cm, at one to two orders of magnitude less memory at matched resolution. With our query integrated into a Model Predictive Controller (MPC), simulated runs in laser-scanned forests at up to 9 m/s collide in 0.4 % through dense branches, against 5.9 % with the best voxel grid and 5.5 % with the best mapless baseline at our resolution, whose failures we attribute to voxel resolution and total latency respectively. We deploy the system onboard at two forest sites and fly at commanded speeds of up to 4 m/s.