This study proposes a framework that unifies detection and localization in passive sonar processing. Conventional sonar systems perform target detection and direction of arrival estimation sequentially: an energy detector first determines signal presence, then beamforming estimates the bearing of maximum energy. This structure separates detection from localization, reducing efficiency in dynamic environments. To address this, we use features capturing spectral and directional characteristics of the acoustic scene. Their discriminative power is verified by evaluating the separability of signal-present and signal-absent conditions in a UMAP-embedded space, confirming their effectiveness for integrated detection and localization. The goal is to build a probabilistic model that estimates the likelihood of target presence across bearing angle. This enables direct inference of both the existence and direction of a signal, offering a robust alternative to threshold-based processing. Work partly supported by KRIT grant funded by the Korea Government (DAPA) (KRIT-CT-23-035, Multi AUV operation Technology for Mine Detection ('23∼'28)) (0.4), and partly supported by Korea Institute for Advancement of Technology (KIAT) Grant funded by the Korea Government (MOTIE) (RS-2023-KI002688, HRD Program for Industrial Innovation) (0.3), and a part of the project titled “Fostering Talent in Advanced Ship Blue Tech (RS-2025-02221147),” funded by the Ministry of Oceans and Fisheries, Korea (0.3).
Kang et al. (Wed,) studied this question.