Development of Optimized Compressive Sensing Algorithms for Wireless Body Area Network

Abstract

The world-wide escalating medical problems and associated diseases bring challenges in preventive newlineand post-surgery care, and need for accurate bio-monitoring systems. Amongst them, heart problems newlinetend to increase globally in an exponential way and demand non-invasive, perpetual ambulatory newlinemonitoring of vital signs of a person to avoid casualties. Conventionally, an electrocardiogram (ECG) newlinehas been used to assess the type of Cardio Vascular Disease (CVD) by medical experts, which may newlinefail to give clear information about peripheral blood circulation and blockages in the coronary arteries. newlineTo overcome these limitations, one more physiological signal such as Photoplethesmogram (PPG) newlinealong with ECG can be employed for detecting the respiratory and cardiovascular system and the newlinecorresponding ailments. newlineRecently the demand for light weight wearable products to obtain a person s vital parameters newlineand transmission of the information is on rise due to increase in health issues across all the age groups. newlineIn this context, energy management in resource-constrained wearable tiny nodes during and after newlinesignal acquisition stages is a critical design aspect especially in the medical Wireless Body Area newlineNetwork (WBAN). Remote Health Monitoring (RHM) systems enabled by WBAN offer a low cost newlinesolution to collect the data of individual patients where wearable bio-sensor devices need to execute newlinelow complexity embedded signal processing algorithms. Amongst the various dimension reduction newlinetechniques developed, the emergent Compressive sensing (CS) algorithm has been found to be most newlineenergy-efficient for encoding biomedical data on resource-constrained sensor nodes. newlineThe main objective of this thesis is to develop optimized compressive sensing newlinealgorithms for WBAN within the CS framework. Many classical CS algorithms have been proposed newlinebut are neither energy-efficient nor achieve higher reconstruction quality of data which comply with newlineclinical acceptance standards. Even though efficient signal communication and detection is