Robust estimation of body temperature based on clustering

Abstract

This paper proposes a method for estimating the body temperature using a temperature sensor installed on the inside surface of a blood-pressure cuff. It is known that the skin surface temperature changes with the outside temperature so that the human body temperature is kept constant. In this paper, based on this property, the body temperature is estimated from the skin surface temperature and the outside temperature measured concurrently while the blood pressure is measured. Considering the actual situation, the estimation could be difficult when the temperature of the blood-pressure meter cuff is greatly different from that of upper arm skin, for example. To solve the problem, we classify in advance various learning samples into multiple classes with different sets of parameters to estimate the body temperature.