engleski

FPGA based two-channel ECG sensor node for wearable applications

Emergence and development of pervasive healthcare applications demand new solutions for continuous and long-term individual’ s health condition monitoring and harmful event detection or, if possible, prediction. Such solutions include, among others, development of wearable, low power and intelligent devices capable of real-time decision making based on physiological parameter analysis. In recent years a large number of wearable devices were designed solely for the purposes of data acquisition and wireless transmission with intent that any further analysis would be done off-line. Prototypes of wearable devices capable for real time on-body physiological parameters processing are at the present time still scarce. Appearance of low power, mixed signal and flash based FPGAs presents new possibilities in development of small, wearable, intelligent and low power real-time biological signal processors used for physiological parameter extraction. Amongst many physiological signals ECG is the key parameter for evaluating an individual’ s well being. The ECG contains vast amounts of data closely related to a person’ s health condition and as such demands broad analysis. Given the complexity of the signal, valuable analysis in many applications can be performed only by using great computational power and professional knowledge. Efficient implementation of data extraction and harmful event detection algorithms from the ECG on mobile platforms is one of the most complex and important areas of research and development in pervasive healthcare applications. To meet the needs of mobile physiological parameters acquisition and analysis we are developing a network of generic wearable FPGA sensor nodes. One of such nodes is a two-channel ECG processor designed for long term real-time analysis and harmful event detection. The fore mentioned ECG node is presented in this paper.

wearable sensor node; real time on-body processing; low power FPGA; two-channel ECG

nije evidentirano