CaMeRL: Collision-Aware and Memory-Enhanced Reinforcement Learning for UAV Navigation in Multi-Scale Obstacle Environments
This work addresses the overlooked problem of obstacle scale variation in UAV navigation, improving performance in environments with both small and large obstacles.
CaMeRL introduces a reinforcement learning framework for UAV navigation that uses collision-aware latent representations and temporal memory to handle multi-scale obstacles, achieving success rate gains of 0.48 and 0.28 in ultra-small and extra-large obstacle settings, respectively, over state-of-the-art baselines.
In obstacle avoidance navigation of unmanned aerial vehicles (UAVs), variations in obstacle scale have received strangely less attention than obstacle number or density. Existing methods typically extract purely geometric features from single-frame depth observations. Such representations tend to neglect small obstacles and lose spatial context under occlusions caused by large obstacles, leading to noticeable degradation in environments with multi-scale obstacles. To address this issue, we propose CaMeRL, a Collision-aware and Memory-enhanced Reinforcement Learning framework for UAV navigation. The collision-aware latent representation encodes risk-sensitive depth cues to preserve fine-grained obstacle structures, thereby improving sensitivity to small obstacles. The temporal memory module integrates observations across frames, mitigating partial observability caused by large-obstacle occlusions. We evaluate CaMeRL with multi-scale obstacles, including ultra-small and extra-large obstacle settings. Results show that CaMeRL outperforms state-of-the-art baselines across all scales, with success rate gains of 0.48 and 0.28 in the ultra-small and extra-large settings, respectively. More importantly, CaMeRL achieves reliable navigation in cluttered outdoor environments.