ECG Signal Based Power Aware System for Obstructive Sleep Apnea Detection

The paper discusses the issues involved in designing a ECG signal based Obstructive Sleep Apnea (OSA)Detection system having low power features and improved battery life. The ECG signal parameters required for the analysis in order to detect OSA are gathered from QRS detection process, following Pan Tompkins algorithm. Heart rate is computed from the data gathered and OSA detection process is implemented with the algorithm providing decision as either normal case without any apnea or presence of Obstructive Sleep Apnea. The algorithm is implemented using MSP430 based MSP430FR6989 launch pad having robust low power and high performance features along with the profiling of power and energy consumption of the entire system. We could improve the code in succession by harnessing the ultra low power capabilities of the controller making use of the Tool features like ULP advisor and Energy trace of Code composer Studio(CCS). The use of Low Power Mode 3 (LPM3) of MSP430 device helped us develop the improved power efficiency of the application code for OSA detection system as compared with the use of other power down modes. Also the life of AA battery is drastically improved for the code using LPM3 as compared with that of CR2032 battery. Here we present the profiles of power and energy in the form of graphs and tabulated data for comparison, using which we have developed a power aware system for OSA detection.