In residential buildings, noise generated by residents or home appliances propagates through building structure and annoys residents on other floors. It is difficult for human to precisely identify the type/position of inter-floor noise, and some conflicts between residents have originated from this incorrect estimation of type/position. Correctly identifying the noise type/position is considered to be the first step in solving noise problem.

We built three different inter-floor noise datasets to study this problem.

A noise signal over a single microphone with a sufficient time duration might contain the dispersive nature of the plate wave or unidentified features. We used data-driven approach to catch these features and identify the noise signal.

• IWAENC 2018 (repo.)
  • In this work, an method for inter-floor noise type/position classification was proposed and validated against an inter-floor noise dataset.
  • We built an inter-floor noise dataset SNU-B36-50 in an office building.
• Appl. Sci. (repo.)
  • This paper is expanded version of IWAENC 2018.
  • Type/position classification of noise at unlearned positions was shown.
  • We collected inter-floor noise at new positions around the noise source positions in SNU-B36-50 and the new data was not used in learning.
• EUSIPCO 2020 (accepted, repo.)
  • We collected more inter-floor noises in two apartment buildings.
  • The generalizability of the proposed method in Appl. Sci. was addressed against the new inter-floor noise datasets.
  • The viability of inter-floor noise type classification knowledge transfer was demonstrated.
• Mobile phone application (under development, repo.)
  • The single acoustical sensor approach can be implemented on a smartphone device.
  • Our inter-floor noise identification method was implemented on a smartphone application (iOS).