waterFSA: A Contact-Less Water Flow Source Analyzer for the Household to Enable HAR and ADL Recognition

The development of a contact-less monitoring system detecting water flow in the household and identifying the objects utilizing it, can provide support for Human Activities Recognition (HAR) and for Activities of Daily Living (ADL). This makes it particularly valuable in the context of Ambient Assisted Living (AAL). This study presents the results of an approach based on analyzing the spectrum of sounds captured in the environment. Two non-invasive solutions are examined, both utilizing a platform with a microphone placed in the environment to detect changes in sound caused by water flow noise. One method is based on spectrum analysis; the other on spectrogram analysis. We present the experiments conducted to test the two approaches in different environments. The two approaches differ in computational cost and are suitable for environments with multiple water sources, requiring no direct contact with plumbing. This makes them particularly suitable for applications related to the use of the bathroom. In particular, the spectrogram approach achieves an accuracy above 90%.