Audio and depth improve robot navigation in dark industrial sites
ADAPT: Audio-Depth Action Planning with Transition-Aware Rewards for Lights-Out Industrial Navigation
Artificial Intelligence
Summary
Robots that need to move around in dark factories cannot rely on normal cameras because there is no light. The authors created a way for robots to use sounds and depth sensors to find their way without any visible light. They also made a new testing set with real factory scenes and different sound conditions to check how well their system works. Their new method, ADAPT, helps robots learn better navigation by understanding sound direction and how to move efficiently, even when it’s noisy. Tests show this approach works well in new places and difficult sound environments.
What this means in practice
- •For industrial automation teams: Enable navigation of robots in completely dark industrial environments using audio and depth sensing to avoid obstacles and reach targets reliably.
- •For robotics system integrators: Implement a reinforcement learning framework that improves robot movement decisions based on sound and depth data in noisy, unlit indoor settings.
Authors
Wenjun Huang, Shenghao Fu, Yian Jin, Yang Ni, Yifan Li, Lichi Li, Xingjian Diao, Pengfei Zhang, Yutong Song, Hanning Chen, Jiarui Cui, Tianyu Yang, Shilong Li, Yu Kong, Mohsen Imani
Abstract
Vision-language-action models for embodied navigation typically rely on RGB observations, limiting their applicability in lights-out industrial environments where visible illumination may be unavailable. We study illumination-independent navigation using spatial audio for target localization and depth for obstacle-aware geometry. To support this setting, we introduce DarkNav-Bench, a benchmark with 24 industrial scenes and four progressively challenging acoustic conditions. We further propose ADAPT, an audio-depth action-planning framework equipped with spatial-audio perception and optimized for closed-loop navigation through reinforcement learning (RL). Our key RL contribution, transition-aware reward shaping (TARS), defines a pose-aware shortest-action potential over a position-heading graph. Unlike conventional geodesic-progress rewards, TARS accounts for both feasible movements and in-place rotations, providing action-consistent feedback for policy optimization. Experiments show that ADAPT consistently outperforms training-based navigation baselines and zero-shot multimodal models on both seen and unseen scenes. On unseen environments, it achieves 90.6% success under Clean conditions and retains 60.0% success under the most challenging Localized Noise setting. These results demonstrate the potential of spatial audio and depth as complementary modalities for reliable RGB-free navigation in lights-out industrial environments.