RiskFly: Frustum-Aligned Spatio-Temporal Risk Fields for One-Stage Agile Flight in Dynamic Clutter

2026-10-03

AI Summary

This study addresses the degradation of dynamic obstacle avoidance into passive reactivity in single-stage planners caused by the absence of spatiotemporal structural signals. To overcome this limitation, we propose RiskFly, which introduces a frustum-aligned inverted spherical range map and a self-predicted spatiotemporal risk field, unifying representation supervision and planning gradients within a shared space. By integrating dual-stream observation, differentiable trajectory querying, and quintic polynomial generation, RiskFly enables end-to-end, mapless agile flight without requiring privileged information. Simulation and zero-shot real-world flight experiments demonstrate that RiskFly significantly improves obstacle avoidance success rates while maintaining low latency on resource-constrained platforms.

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.

Problem

• Agile flight
• Dynamic clutter
• One-stage planner
• Spatio-temporal risk
• Reactive maneuvers

Innovation

• frustum-aligned risk field
• one-stage planner
• spatio-temporal risk prediction
• differentiable trajectory query
• agile flight