Wireless Body Area Network WBAN for Real Time Wearable Physiological parameters Monitoring and Algorithm to Classify Health Status

Abstract

The recent advancement in the wireless communication technology has given way for newlinethe development of Wireless Body Area Network (WBAN). In wireless body area newlinenetwork an array of sensors are strategically placed on the surface of the body or even newlineimplanted under the skin surface. These sensors are placed inside clothing, wrist newlinebands, chest belts etc. This is the emerging technology for the wearable physiological newlinemonitoring applications. It has revolutionized the computer communication with the newlinesmall input output devices which can work as an independent system to acquire newlinephysiological data from the body and transmit them wirelessly. The advantage of newlinewearable sensors is that, these sensors are comfortable to wear and they can be used newlinefor longer duration monitoring of vital signals. In this work three types of wearable newlinesensors are designed using dry electrodes and they are a) textile electrode b) surface newlineelectrode and c) needle electrode. newlineIn the proposed work the wearable system is designed using three sensor nodes and a newlinesink node. The sensor node-1 is designed for acquiring electrocardiogram signal and newlinebody temperature, sensor node-2 is designed to acquire electroencephalogram and newlinesensor node-3 is for acquiring galvanic skin response and photoplethysmography newlinesignals. The data from the sensor node to the sink node is transferred using ZigBee newlinecommunication and the wireless body area network is implemented in a star topology. newlineIn this study different filtering techniques to remove powerline interference and newlinebaseline wandering noise in electrocardiogram signal is carried out. Finite impulse newlinefiltering, notch filtering and discrete wavelet transform techniques are being used and newlinea comparative analysis is done. For baseline wandering removal an algorithm using newlineDaubechies wavelet of tenth order is used along with moving average filtering. newlineFurther in this study for non-invasive blood pressure measurement different machine newlinelearning regression techniques like decision tree, ensemble tree method, linear newlinereg