Hidden Markov Model energy conservation approach for continuous monitoring of vital signs in geriatric care applications

Abstract

In the recent healthcare crisis engendered by the Covid19 pandemic, wireless body area networking devices have started to play a significant role in mitigating the health problems of the elderly. The energy conservation of the device during the temporary disconnection of the sensor node can play a vital role in the broader acceptance of this technology. Here, a probabilistic hidden Markov model (HMM) is used for energy conservation, a relatively less explored area of energy conservation approach within the field of wireless sensor networks. Since the vital signs of heart rate and blood pressure are highly correlated, the heart rate and blood pressure readings are taken for model development. The classification of normal and critical data is based on the probability of observation sequences in the particular model. The hidden states are estimated using the observation sequence and the HMM parameters during disconnection. Accuracy between 0.9 and 1.0 is obtained for different series. Dynamic threshold limits are included for more adaptability of the model for varying physiological conditions of the patients. The energy conservation possible using the model is discussed. This model presents a novel approach to energy conservation using HMM, which will help continuous home monitoring of vital parameters in geriatrics.