Indoor localization method with environmental sound for human behavior measurement

2024 Fiscal Year Final Research Report

• Project/Area Number: 22K17922
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 61010:Perceptual information processing-related
• Research Institution: National Institute of Advanced Industrial Science and Technology
• Principal Investigator: Ogiso Satoki 国立研究開発法人産業技術総合研究所, 情報・人間工学領域, 研究員 (10821738)
• Project Period (FY): 2022-04-01 – 2025-03-31
• Keywords: 環境音 / 屋内測位

Outline of Final Research Achievements

This study aims to establish a positioning method that can estimate two-dimensional coordinates using a single microphone by utilizing the knowledge of environmental sounds and sound maps. The proposed method estimates the amplitude of each sound source using non-negative matrix factorization on the spectrogram of the sound, and then calculates the position likelihood by comparing it with the sound map showing the amplitude of each sound source at each location. The proposed method was tested within the research institute and in operational service areas to evaluate positioning accuracy. During this process, it was suggested that information about the absence of sound contributes to positioning. Through this study, the proposed method showed the usefulness of environmental sounds as a source of information for obtaining absolute positions without the need for device installation.

Free Research Field

屋内位置情報，音響計測

Academic Significance and Societal Importance of the Research Achievements

本研究は環境音のみを用いて位置推定を行う方法を示すものであり，従来は部屋の識別などまでしかできなかった環境音で位置座標を推定する方法を示した．また，実験を通じて提案法の特性を明らかにしており，事前に音の知識を用いることで少なくとも10 dB程度雑音に対し頑健にできること，営業中のサービスエリア内で冷蔵庫等の音を用いて位置推定が可能であることを示した．また，小さい音や音が聞こえないことが位置推定に重要なことを示唆した．これらはマイクさえあれば概ね位置がわかることを示しており，簡便な人行動計測や小型IoT機器などの位置情報計測に貢献する．