Low power wearable wireless ECG system for long-term homecare
Aachen (2016, 2017) [Dissertation / PhD Thesis]
Page(s): 1 Online-Ressource (xxii, 116 Seiten) : Illustrationen, Diagramme
This thesis has proposed a novel wearable wireless ECG system. With the consideration of long-term homecare application, it strives to control the size and power consumption of the sensor node. As a result, this thesis is devoted in three aspects: new electrode placements design, wireless ECG system design and multiple power control technologies.In the new electrode placements investigation, an experiment was designed to investigate the best limb electrode placements. The experiment compared 14 different placements for limb electrodes. The detected signals of different placements were compared with the standard lead system. The best placements for four limb electrodes were selected according to the correlation coefficients between the standard and new placements.In the wireless ECG system design, a low noise analog frontend was implemented for the ECG signals, considering practical issues like dc offset caused by body motion, EMI coupled from the power line and electrode impedance mismatch. The measurement result showed excellent performance even under body motion. With the new electrode placements and the low noise analog front end, two wireless ECG systems were implemented with ZigBee and BLE. The sizes of both sensor nodes were controlled in 5.5 cm x 2.5 cm, with which the sensor node was able to be conveniently worn on the body without affecting user's mobility. The ECG signals were displayed on PC or smartphone in real time.This work applied multiple power control technologies in both analog and digital ways to extend the battery life. Firstly, adjustable power mode control was operated in the ZigBee and BLE sensor nodes with battery lives of 52 hours and 55hours respectively. Secondly, dynamic transmission power control in ZigBee system was utilized to adjust the Tx output power dynamically according to the received signal strength indicator. It saved 20% - 30% power during regular movements. Thirdly, Compressed Sensing (CS) was applied to reduce the size of the transmitted data. Digital CS was firstly implemented in the BLE system. ECG signals were successfully reconstructed in real time with little distortion under compression ratio of 2. The battery life was extended by 12 hours. Analog CS was also implemented by an integrated encoder using 0.13 μm CMOS technology. Instead of using 64 parallel SAR-ADCs, only one SAR-ADC was employed. The total average power consumption was 23.5 μW. Finally, an integrated analog front end was designed and implemented in 0.13 μm CMOS technology. The offset and 1/f noise of the first Gm were noted. Ac coupling circuit and chopped current coupled instrumentation amplifier were the solutions to reduce the noises appeared in ECG signals. The measurement results showed that, the chip only consumed 5 μA current and supplied 46.3 dB gain and 0.56 Hz – 90 Hz bandwidth. Furthermore, high range dc-electrode offset (500 mV) and common-mode voltage (0.2 V - 1.0 V) were achieved to provide high tolerance for body motion and electrode mismatch.