Audio localization for drones adapts timing to speed detection
Audio-based UAV Localization with Adaptive Temporal Correspondence via Reinforcement Learning
RoboticsSound
Summary
Detecting drones by sound is a cheap way to spot them even when it's dark. Existing methods look at fixed lengths of sound, which can either take too long or miss details. The authors developed a system that listens to variable lengths of sound based on how reliable the signal is, decided using a kind of trial-and-error learning. This method helps find where drones are in 3D more quickly and still accurately. They tested it in many situations and it worked well across different conditions.
What this means in practice
- •For security teams: Use adaptable audio analysis to locate drones quickly for early detection in restricted areas.
- •For wildlife monitoring teams: Implement adaptive sound-based localization to track airborne objects in low-visibility environments without relying on light.
Authors
Haoxiang Lei, Mingzheng Feng, Daotong Wang, Shenghai Yuan
Abstract
Audio-based localization provides a low-cost and illumination-independent sensing solution for anti-UAV early warning. However, existing methods typically rely on a predefined fixed audio segment length, which limits temporal correspondence and creates a trade-off between sufficient acoustic evidence and timely localization. To address this issue, we propose an audio-based localization framework with adaptive temporal correspondence. A probe segment is first used to extract a compact acoustic state that characterizes the reliability and consistency of the observation. Guided by the state, a reinforcement learning controller dynamically determines the required audio window size for each localization decision. The selected audio segment is then processed by a Mamba-based localization network with adaptive temporal feature modulation for 3D position estimation. Extensive experiments demonstrate that our method achieves competitive 3D localization accuracy with substantially reduced temporal correspondence latency compared to SOTA methods and exhibits strong generalization across scenarios.