LNN-Fly: Continuous-Time UAV Navigation for Robust Obstacle Avoidance under Timing Mismatch

2026-06-27

AI Summary

This work addresses the degradation of obstacle avoidance performance in real-world drone deployment caused by sim-to-real discrepancies, sparse perception, and varying control frequencies. To this end, the authors propose a deployment-oriented, continuous-time navigation policy that integrates a dynamic programming–inspired structured recurrent update, explicitly models the control interval Δt, and introduces an input-driven adaptive forgetting gate to refresh stale hidden states in hazardous regions—balancing safety and state consistency. By training within a differentiable simulator that incorporates deployment perturbations, this approach uniquely combines explicit temporal modeling with adaptive forgetting, significantly enhancing robustness to perception delays, control jitter, and frequency shifts. Experiments demonstrate substantially improved simulated avoidance performance, 100% success across 20 cross-frequency indoor flights on a real quadrotor, and low inference latency (median 0.514 ms on a desktop GPU, ~2.5 ms on an onboard CPU, P95 < 30 ms), enabling zero-shot transfer.

Abstract

End-to-end unmanned aerial vehicle (UAV) navigation can achieve impressive agility in simulation, yet its obstacle-avoidance behavior often degrades after deployment because the policy must tolerate simulator mismatch, sensing irregularity, and variable-rate control. These effects are especially dangerous in cluttered environments, where stale observations or short control irregularities can directly lead to collisions. We present LNN-Fly, a deployment-oriented continuous-time navigation policy for LiDAR-based UAV obstacle avoidance. The policy combines a dynamic-programming-inspired structured recurrent update, explicit conditioning on the elapsed control interval Δt, and an input-driven adaptive forgetting gate that refreshes stale latent state near hazards while preserving consistency during sustained maneuvers. It is trained with differentiable rollouts that incorporate deployment-relevant sensing and timing perturbations. In simulation, LNN-Fly improves obstacle-avoidance performance in the tested settings and shows better tolerance to reduced control frequency, sparse observations, and control-period jitter. It also transfers zero-shot from a simplified differentiable simulator to a physical quadrotor. In indoor cross-frequency real-world tests, the system achieves 100% success over 20 flights, while policy inference has a median latency of 0.514 ms on a desktop graphics processing unit (GPU) and about 2.5 ms on the onboard central processing unit (CPU), with onboard P95 latency below 30 ms.

Problem

• timing mismatch
• obstacle avoidance
• UAV navigation
• simulator-to-reality gap
• control irregularity

Innovation

• continuous-time navigation
• timing mismatch robustness
• adaptive forgetting gate
• differentiable rollout
• zero-shot sim-to-real transfer