LOCUS: LOcalization with Channel Uncertainty and Sporadic Energy

TL;DR

LOCUS is a deep learning framework designed to improve sound source localization accuracy in energy-harvesting systems with missing data by reconstructing corrupted features and restoring multichannel information.

Contribution

It introduces a novel multi-module approach combining feature corruption detection, reconstruction, and data restoration to enhance localization performance under energy constraints.

Findings

Achieves up to 36.91% error reduction on benchmark datasets.

Improves real-world DoA accuracy by 25.87-59.46%.

Provides a new dataset for localization research under energy harvesting conditions.

Abstract

Accurate sound source localization (SSL), such as direction-of-arrival (DoA) estimation, relies on consistent multichannel data. However, batteryless systems often suffer from missing data due to the stochastic nature of energy harvesting, degrading localization performance. We propose LOCUS, a deep learning framework that recovers corrupted features in such settings. LOCUS integrates three modules: (1) Information-Weighted Focus (InFo) to identify corrupted regions, (2) Latent Feature Synthesizer (LaFS) to reconstruct missing features, and (3) Guided Replacement (GRep) to restore data without altering valid inputs. LOCUS significantly improves DoA accuracy under missing-channel conditions, achieving up to 36.91% error reduction on DCASE and LargeSet, and 25.87-59.46% gains in real-world deployments. We release a 50-hour multichannel dataset to support future research on localization under energy constraints. Our code and data are available at: https://bashlab.github.io/locus_project/