RiskFly plans safer drone flights with spatio-temporal risk prediction
RiskFly: Frustum-Aligned Spatio-Temporal Risk Fields for One-Stage Agile Flight in Dynamic Clutter
Robotics
Summary
Flying drones quickly through cluttered and changing spaces is tricky because the drone needs to know not just where obstacles are but also when they might be dangerous. The authors created RiskFly, a system that predicts danger in both space and time, matching the drone’s planned path to avoid hazards better. By training the system to predict a detailed risk map rather than just a single safety score, flights become smoother and less reactive. Their tests show it works well even without detailed maps, directly from the drone’s sensors.
What this means in practice
- •For drone developers: Enable drones to plan and execute fast, safe flights through unknown and moving obstacle environments using onboard depth sensors without relying on maps.$Commercial implications: This method allows building commercial drones for inspection and delivery that navigate dynamic urban or industrial spaces more reliably.
- •For robotic vehicle engineers: Build agile ground or aerial robots that predict risk along planned paths in cluttered, moving environments to improve collision avoidance at runtime.
Authors
Luxia Ai, Haopeng Chen, Yuchao Mei, Guohao Zhang, Wenbing Tao
Abstract
Agile flight in unknown, cluttered, and dynamic environments requires a planner that knows where and when danger will appear, not only that a trajectory is dangerous. One-stage learning-based planners trained with differentiable privileged costs are fast and expert-free, but the only signal reaching their encoder is a scalar trajectory cost with no spatial or temporal structure, so avoidance degrades into late reactive maneuvers. We present RiskFly, a one-stage planner that predicts risk in the same space in which it acts. A dual-stream observation pairs a short depth sequence with a frustum-aligned inverted spherical range-map sequence, whose angular cells match the end-state proposals one to one. An auxiliary head regresses a frustum-aligned spatio-temporal risk field, supervised by a privileged closest-point-of-approach (CPA) target. This self-predicted field is queried differentiably along the instantiated quintic trajectory at its own arrival times, and also enters the training objective, so representation supervision and planning gradients meet in a single space. Privileged signals are discarded at deployment, and the planner runs map-free from onboard depth and proprioception. Extensive simulation and zero-shot real-world flights on resource-constrained platforms show higher success rates at comparable end-to-end latency.