Sensing method and device

Abstract

The application discloses a method for determining the status of a subject, such as the rate of fluid flow in a pipe, whether a pipe is leaking, or whether water in a pipe is at risk of developing Legionella bacteria. Sensors measure parameters 210 such as temperature, humidity, pressure, light, motion, or audio. A parameter-specific machine learning model 220 processes a respective parameter to derive a parameter-specific indication 230 of the subject’s status. For example, the machine learning model may use a temperature measurement to classify a pipe as leaking, having normal flow, or having no flow. A multi-parameter model 240 considers this determination alongside another measurement to derive a multi-parameter indication 250 of the subject’s status. For example, the flow classification derived from temperature may be considered alongside an audio measurement to derive a likelihood that the pipe is leaking. A signal is then output based on the multi-parameter indication. For example, a light may illuminate if the pipe is leaking. One sensor may wake other sensors from an inactive state if its measurement meets a criterion. Other subjects include an environment, a living organism, a patient, or a body part.

Images