A μw Bio-impedance Sensor with 276 μs Settling Time for Portable Blood Pressure Monitoring System

Transcription

1 JOURNAL OF SEMICONDUCTOR TECHNOLOGY AND SCIENCE, VOL.17, NO.6, DECEMBER, 2017 ISSN(Print) ISSN(Online) A μw Bio-impedance Sensor with 276 μs Settling Time for Portable Blood Pressure Monitoring System Kwantae Kim 1, Minseo Kim 1, Hyunwoo Cho 2, Kwonjoon Lee 1,2, and Hoi-Jun Yoo 1 Abstract An ultra-low power bio-impedance sensor for wrist watch-type blood pressure monitoring system is implemented in 0.18 μm CMOS technology under 1.8 V supply. The duty-cycled current generator (DCG) enables low power and high amplitude current injection with 93.58% power reduction in a current generator block. A fast-settling readout circuit with DC balanced amplifier reduces settling time to 276 μs, allowing bio-impedance signal acquisition even when the motion artifact induced circuit transient occurs. The simulated mm² single chip bio-impedance sensor consumes only μw. Index Terms Arterial pulse wave (APW), ambulatory blood pressure monitoring, bioimpedance, duty-cycled current generator (DCG), pulse transit time (PTT) I. INTRODUCTION According to recent statistics in the US, cardiovascular disease (CVD) is the most significant cause of mortality rate among humans [1]. Among CVDs, hypertension is considered as a major risk factor with potentials of sudden deaths. For this reason, the demands of ambulatory blood pressure (BP) monitoring devices are increasing recently. But conventional cuff-based blood pressure monitoring devices are occlusive and noncontinuous, resulting in cumbersome and uncomfortable blood pressure measurements. Furthermore, practical requirements of fifteen to thirty minutes of resting time made the conventional cuff-based devices monitor only short-term regulations of blood pressure [2]. To address this problem, a blood pressure monitoring method based on pulse transit time (PTT) measurement has emerged. This method measures the time difference between the R peak of the ECG and the maximum slope of the arterial pulse wave (APW), which is called PTT as shown in Fig. 1. The linear calibration of PTT can be used to estimate blood pressure [3]. Since the measuring ECG and APW signals does not require occlusive and minutes long periods of cuffs, the PTT-based blood pressure monitoring method is non-occlusive, continuous. There are two widely used methods of APW measurements. A light-based APW measurement as known as photoplethysmography (PPG) [4], emits LED light to blood vessels, and measures reflected LED light from hemoglobin. The measured light intensity varies with blood flow which is APW signal. However, this method typically consumes large power, due to tens of Manuscript received Nov. 12, 2017; accepted Nov. 20, School of EE, KAIST, 291 Daehak-ro, Yuseong-gu, Daejeon, Korea 2 Healthrian, 71 Jukdong-ro, Yuseong-gu, Daejeon, Korea kwantae.kim@kaist.ac.kr Fig. 1. Bio-impedance measurement for PTT calculation.

2 JOURNAL OF SEMICONDUCTOR TECHNOLOGY AND SCIENCE, VOL.17, NO.6, DECEMBER, Fig. 2. Proposed bio-impedance sensor architecture. ma scale current consumption and high level of DC bias voltage which is required to operate LED. To overcome above problems, impedance-based APW measurement (bio-impedance, Bio-Z) have been used, which injects tens of khz range AC current to human body. When the current is injected, blood flow synchronized AC impedance (Z AC in Fig. 1) makes an amplitude modulated voltage waveform, which can be demodulated to the baseband. Because the bio-impedance sensors do not require large power consuming LED, it has an advantage over PPG for ambulatory blood pressure monitoring devices. Although the bio-impedance sensor consumes less power than the PPG method, but it still consumes too much power (exceeding 100 μw) to be used in ambulatory blood pressure monitoring devices. In such previous bio-impedance sensors [5-7], assuming 10 Mbps data rates [8] and 12 mah (5V) battry, the operation time is less than 30 days. The most power consuming building block was the current generator. There are two main reasons of it. First, the requirement of over 10 khz current injection. Due to the limited input voltage ranges of readout circuit, large values of electrode-tissue contact impedance [9] results in overall circuit saturation. Second, the requirement of high amplitude current injection. Since the AC impedance of human wrist is in the tens of mω scale, at least hundreds of μa of current injection is required to obtain amplitude modulated input signal in μv scale. This paper proposes a duty-cycled current generator (DCG) for bio-impedance sensor IC [10] to reduce the average amount of power consumption, without compromising amplitude or frequency of current injection. The selection of duty cycle is based on Association for the Advancement of Medical Instrumentation (AAMI) standards satisfaction capability, and the core current generator block adopts low power pseudo-sine synthesizer [11]. In addition, the proposed DC balanced amplifier shortens the settling time when circuit transients due to the motion artifact of users. It opens low impedance path to increase settling current in high pass filter, obtaining fast-settling bio-impedance signal acquisition to readily monitor the long-term regulations of blood pressure. The rest of this paper is organized as follows. In Section II, the overall architecture of the proposed bioimpedance sensor will be described. The detailed operation flow and circuit implementations of DCG and fast-settling readout are then presented in Section III. In Section IV, the simulation results will be shown and Section V concludes the paper. II. ARCHITECTURE Fig. 2 shows the proposed architecture of bioimpedance sensor. The sensor chip is composed of two main blocks: First, DCG injects duty-cycled 100 khz and 100 μapp pseudo-sine current to the human wrist. The core current generator is based on low power pseudo-sine synthesizer, which does not require power consuming pure-sine oscillator [5]. In addition, 8% duty-cycled operation achieved further power reduction in DCG. Duty cycle operation results in an additional PTT error (ΔPTT) of less than 163 μs, which can still meet the AAMI standard requirements with less than 0.1 mmhg of additional mean arterial pressure estimation error (ΔMAP). Second, the fast-settling readout circuit includes a proposed DC balanced amplifier. The DC balanced amplifier is controlled by a saturation detector. It sends an enable signal to DC balanced amplifier when the output node of the programmable amplifier (PGA) is saturated. Furthermore, the duty cycle synchronized comparator achieved power reduction by operating only when the current is being injected. III. CIRCUIT DESCRIPTION 1. Duty-cycled Current Generator (DCG) The operating principle of DCG is described in Fig. 3.

3 914 KWANTAE KIM et al : A m W ΜW BIO-IMPEDANCE SENSOR WITH 276 m s SETTLING TIME FOR PORTABLE BLOOD (a) Fig. 3. Upper bounded PTT measurement error (ΔPTT) due to duty cycling. It adopts duty cycling to reduce the average power consumption of current generator. According to repeated ON/OFF periods, the bio-impedance measurement is sparsely conducted. Then the resulting amplitude modulated waveform is shaped like several superposed step functions. Note that the envelope of superposed step functions is APW signal. Since detecting the maximum slope of APW is required to extract PTT, an error time (t Error ) arisen from the duty-cycled operation affects the measured PTT value. However, this error time (t Error ) is inevitably smaller than the ON/OFF period (t ONOFF ) of the duty cycle. This is because the two points at which the APW maximum slope occurs will be within t ONOFF. This means that t Error can not be greater than t ONOFF and there is a maximum of additional PTT errors (ΔPTT) due to duty cycle. Therefore, we can control ΔPTT by adjusting the duty cycle of DCG, reducing the average power consumption of DCG along with the allowable ΔPTT value. Recently, a linear fitting Eq. (1) was found between the PTT and MAP, satisfying Association for the Advancement of Medical Instrumentation (AAMI) standards for the ambulatory blood pressure devices [3]. In this work, we selected the duty cycle of DCG based on (1) and calibrated further by trial and error method. MAP = PTT (1) Although (1) is not the golden equation which is applicable for every person, however, the implemented DCG can be easily adjusted and redesigned, according to the golden equation which will be found in future research of PTT-based blood pressure monitoring (b) Fig. 4. (a) ΔPTT selection for target ΔMAP, (b) selection of duty cycle. method. Fig. 4 show the procedures of how the duty cycle value can be chosen. First, we selected allowable ΔPTT which results in allowable ΔMAP, according to (1). ΔMAP of less than 0.1 mmhg does not affect the implemented blood pressure monitoring device because the AAMI standard requires an estimation error of less than 5 mmhg on average and the standard deviation of no more than 8 mmhg. Therefore, the maximum selected ΔPTT is set to less than 163 μs according to (1) and can only result in a ΔMAP of < 0.1 mmhg. Note that the designed ADC operates by 100 khz sampling rate to obtain 10 μs of sampling intervals, enabling the Bio-Z readout circuit to acquire < 163 μs of target ΔPTT value. Second, we estimated the duty cycle by fine-tuning the obtained ΔPTT from the simulated Bio-Z readout, using the trial and error method. Because of the limited bandwidth (50 Hz) of readout circuit, it requires a finite amount of settling time which is induced by ON/OFF operation of DCG, during the large sloping region of APW. This settling time causes additional PTT measurement error that makes ΔPTT value larger, resulting in the need of a trial and error process. Because the required minimum duty cycle to satisfy ΔPTT < 163 μs with 100 khz DCG frequency is 6.13%, we finally

4 JOURNAL OF SEMICONDUCTOR TECHNOLOGY AND SCIENCE, VOL.17, NO.6, DECEMBER, Fig. 6. Operation description of DC balanced amplifier. Fig. 5. Current DAC core of DCG. have chosen 8% of duty cycle resulting in 146 μs of ΔPTT. The core part of DCG includes thermometer-coded current DACs which is more robust to mismatch problem than binary-coded DACs, operating with 16 bits of a sinusoid synthesized lookup table avoiding to use a power consuming pure-sine oscillator. Fig. 5 shows the implemented current DAC core. When the 16 bits of absolute-valued digital sinusoids are sent to current DACs, the output chopper switches the directions of output currents, resulting in complete sinusoid shapes. To isolate DC current flowing from DCG to human body, off-chip CHPF is added for safety, regarding to IEC60601 standards [12]. 2. Fast-settling Readout Since amplitude modulated bio-impedance signal has to be demodulated, DC impedance (Z DC in Fig. 1) results in DC voltage in a demodulated readout. Also, since only the AC impedance is related to the APW signal, a highpass filter with a cut-off frequency of 0.5 Hz or less is required to eliminate the DC impedance. Therefore, abrupt transient response leads to practically 10 seconds of settling time, which is too slow for users to monitor blood pressure. For example, the tension applied by the wrist muscles causes immediate DC impedance changes, resulting in saturation of circuits with high-pass filters. At this time, the settling current flows through the resistor, which is very small due to the large RC time constant of the high-pass filter. To solve this problem, the proposed DC balancing operation increases settling current by quickly shifting cut-off frequency to a higher range. When the PGA saturates, the implemented saturation detector sends an enable signal to the DC balanced amplifier to change its operational phase to a fast-settling phase, during the short time interval. This DC balancing operation is disabled when the saturation detector finds the PGA output node to end the settling, allowing the DC balanced amplifier to restart amplification. The proposed DC balanced amplifier consists of a capacitive feedback amplifier [13] (Same architecture of implemented instrumentation amplifier (IA)), with a capacitor, a pseudo resistor and the simple transmission gate. The operation of the DC balanced amplifier has two phases as shown in Fig. 6. In the amplifying phase, the transmission gate switches OFF, and the amplifier works with a gain of C 1 /C 2. In contrast, in the fast-settling phase, the transmission gate switches ON so that low impedance path can be formed to increase settling current. Because most of the current flows through the transmission gate, the amplifier simply works as unity gain buffer. The implemented saturation detector is shown in Fig. 7. It consists of two comparators and additional logic gates. This circuit detects dynamic range deviation of the PGA output node, and sends an enable signal to DC balanced amplifier. Because the DCG is duty-cycled, the saturation detector can turn on only when the current is being injected. Therefore, the two comparators work in the duty-cycled manner, which reduces the average power consumption. As a result, an average of 4.8 nw of power consumption could be achieved. IV. SIMULATION RESULTS The presented bio-impedance sensor IC is simulated in SMIC 0.18 μm CMOS process. Fig. 8 shows the chip

5 916 KWANTAE KIM et al : A m W ΜW BIO-IMPEDANCE SENSOR WITH 276 m s SETTLING TIME FOR PORTABLE BLOOD Fig. 9. Current injection capability of DCG. Fig. 7. Implemented saturation detector. Fig. 10. Frequency response of DC balanced amplifier. Fig. 8. Layout photograph and performance summary. layout photograph and performance summary, showing the entire area of mm² and μw of total power consumption. Fig. 9 shows the constant current injection capability of DCG. It can inject constant amplitudes of 100 khz pseudo-sine currents, up to kωs of load condition. Fig. 10 shows the frequency response of DC balanced amplifier. It shows a flat band gain close to 0 db in the fast-settling phase. The cut-off frequency increased to 128 khz from Hz, because of the low impedance transmission gate with forms very small RC constant with C 2 (6 pf) capacitor. With this result, the ON state of transmission gate can be calculated as 207 kω. The result of the DC balancing operation is shown in Fig. 11. When the DC balancing operation is enabled, Fig. 11. Waveform of DC balancing operation. settling time (within 1%) is reduced to 151 μs, resulting in fast-settling operation. Since the saturation detector operates in a duty-cycled manner, the total settling time is 276 μs. Fig. 12 compares the output waveform of the PGA, with and without duty cycling of DCG. We used MIMIC II online waveform database [14] of APW. The measured ΔPTT is 146 μs, which means we can guarantee less than 0.1 mmhg of ΔMAP estimation error according to (1). Performance comparison to previous works are summarized in Table 1, with the bio-impedance sensors that achieved over 100 μapp of injection amplitudes which is the requirement to obtain μv range input signal

6 JOURNAL OF SEMICONDUCTOR TECHNOLOGY AND SCIENCE, VOL.17, NO.6, DECEMBER, readout circuit, the transmission gate was added to a capacitive amplifier to make low impedance path in high-pass filter which is controlled by low-power saturation detector. It enables 276 μs of fast-settling operation even with the motion induced circuit transient occurrences. ACKNOWLEDGMENTS Fig. 12. Readout waveform of proposed bio-impedance sensor. Table 1. Performance comparison This work was supported by Institute for Information & communications Technology Promotion (IITP) grant funded by the Korea government (MSIT) (No ,Intelligent Processor Architectures and Application Softwares for CNN (Convolutional Neural Network)-RNN (Recurrent Neural Network)). REFERENCES from human wrist. Comparing to previous works, this work achieved the lowest power consumption without compromising amplitude or frequency of injection current, providing fast-settling functionality for abrupt DC impedance changes. Although the amplitude of injection current is smaller than previous works, this work still depicts less power consumption even when the amplitude of the current is normalized. V. CONCLUSIONS A μw bio-impedance sensor IC with dutycycled pseudo-sine current generator and fast-settling readout with abrupt DC impedance changes for wrist watch-type blood pressure monitoring system is presented. The proposed current generator advances the state-of-the-art bio-impedance sensors by consuming only 5.77 μw without compromising the amplitude or the frequency of the injection current, leading to 93.58% power reduction. For achieving fast-settling operation in [1] D. Mozaffarian et al., Heart disease and stroke statistics 2016 update: a report from the American Heart Association, Circulation, vol. 133, no. 4, pp. e38 e360, [2] T. G. Pickering et al., Recommendations of blood pressure measurement in humans and experimental animals. Part 1: blood pressure measurement in humans. A Statement for Professionals from the Subcommittee of Professional and Public Education of the American Heart Association Council on High Blood Pressure Research, Hypertension, vol. 45, no. 1, pp , [3] J. Solà et al., Noninvasive and nonocclusive blood pressure estimation via a chest sensor, IEEE Transactions on Biomedical Engineering, vol. 60, no. 12, pp , Dec [4] J. Allen et al., Photoplethysmography and its application in clinical physiological measurement, Physiological measurement, vol. 28, no. 3, pp. R1 39, Feb [5] L. Yan et al., A 3.9mW 25-electrode reconfigured sensor for wearable cardiac monitoring system, IEEE Journal of Solid-State Circuits, vol. 46, no. 1, pp , Jan [6] W. Lee and S. Cho, Integrated all electrical pulse wave velocity and respiration sensors using bioimpedance, IEEE Journal of Solid-State Circuits, vol. 50, no. 3, pp , Mar

7 918 KWANTAE KIM et al : A m W ΜW BIO-IMPEDANCE SENSOR WITH 276 m s SETTLING TIME FOR PORTABLE BLOOD [7] C.-C. Tu et al., A 135-μW 0.46-mΩ/ Hz thoracic impedance variance monitor with squarewave current modulation, in Proc. IEEE Asian Solid-State Circuits Conf., Nov. 2014, pp [8] P. Mercier and A. P. Chandrakasan, A 110μW 10Mb/s etextiles Transceiver for Body Area Networks with Remote Battery Power, IEEE Int. Solid-State Circuits Conf. (ISSCC) Dig. Tech. Papers, pp , Feb [9] K. Song, H. Lee, S. Hong, H. Cho, U. Ha, and H.-J. Yoo, A sub-10 na DC-balanced adaptive stimulator IC with multi-modal sensor for compact electro-acupuncture stimulation, IEEE Trans. Biomed. Circuits Syst., vol. 6, no. 6, pp , Dec [10] K. Kim et al., A 54-μW fast-settling arterial pulse wave sensor for wrist watch type system, in Proc. IEEE Int. Symp. Circuits and Systems (ISCAS), May. 2016, pp [11] S. Kim et al., A 20 μw intra-cardiac signalprocessing IC with 82 db bio-impedance measurement dynamic range and analog feature extraction for ventricular fibrillation detection, in Proc. IEEE ISSCC Dig. Tech. Papers, Feb. 2013, pp [12] Medical Electrical Equipment- Part 1: General Requirements for Basic Safety and Essential Performance, International Standard, IEC , [13] R. Harrison and C. Charles, A low-power lownoise CMOS amplifier for neural recording applications, IEEE Journal of Solid-State Circuits, vol. 38, no. 6, pp , Jun [14] A. L. Goldberger et al., Physiobank, physiotoolkit, and physionet components of a new research resource for complex physiologic signals, Circulation, vol. 101, no. 23, pp. e215 e220, Kwantae Kim received the B.S. and M.S. degrees in electrical engineering from the Korea Advanced Institute of Science and Technology, Daejeon, South Korea, in 2015 and 2017, respectively, where he is currently pursuing the Ph.D. degree. During his graduate study, he worked with Healthrian R&D Center, Daejeon, Korea for bio-potential readout IC design. His research interests include low-power bioimpedance sensor IC and digitally-assisted analog circuits. Minseo Kim received the B.S. degree in semiconductor system engineering, Sung Kyun Kwan University (SKKU), Seoul, South Korea, in 2014, and the M.S. degree in electrical engineering from the Korea Advanced Institute of Science and Technology, Daejeon, South Korea, in 2016, where he is currently pursuing the Ph.D. degree. His current research interests include low-power biomedical SoC for wearable healthcare system. Hyunwoo Cho received the B.S. degree from the Department of Physics, Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, and the M.S. and Ph.D. degrees in electrical engineering from KAIST in 2010, 2012, and 2016, respectively. He is currently working as a director of the headquarter of Healthrian, which is mobile healthcare solution company. He has worked on developing the power and energy efficient CMOS wireless transceiver for portable and wearable devices working around the human body. His research interests include lowpower and low energy body area network transceiver design and body channel characteristics analysis.

8 JOURNAL OF SEMICONDUCTOR TECHNOLOGY AND SCIENCE, VOL.17, NO.6, DECEMBER, Kwonjoon Lee received the B.S. degree (summa cum laude) in electrical engineering from the Sung Kyun Kwan University (SKKU), Suwon, Korea, in 2012, and M.S. degree in electrical engineering from the Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, in From 2014 to 2015, he worked at Samsung Electronics for healthcare IC design. He is currently working at Healthrian, Daejeon, Korea. His research interests include system design for wearable healthcare device including application searching, specification setting and bio-medical SoC design. He is also interested in bio-signal processing such as electrocardiogram (ECG), electromyogram (EMG) and arterial pulse wave (APW). Hoi-Jun Yoo graduated from the Electronic Department of Seoul National University, Seoul, Korea, in 1983 and received the M.S. and Ph.D. degrees in electrical engineering from the Korea Advanced Institute of Science and Technology (KAIST), Daejeon, in 1985 and 1988, respectively. Since 1998, he has been the faculty of the Department of Electrical Engineering at KAIST and now is a full professor. From 2001 to 2005, he was the director of Korean System Integration and IP Authoring Research Center (SIPAC). From 2003 to 2005, he was the full time Advisor to Minister of Korea Ministry of Information and Communication and National Project Manager for SoC and Computer. In 2007, he founded System Design Innovation & Application Research Center (SDIA) at KAIST. Since 2010, he has served the general chair of Korean Institute of Next Generation Computing. His current interests are computer vision SoC, body area networks, biomedical devices and circuits. He is a coauthor of DRAM Design (Korea: Hongrung, 1996), High Performance DRAM (Korea: Sigma, 1999), Future Memory: FRAM (Korea: Sigma, 2000), Networks on Chips (Morgan Kaufmann, 2006), Low-Power NoC for High-Performance SoC Design (CRC Press, 2008), Circuits at the Nanoscale (CRC Press, 2009), Embedded Memories for Nano-Scale VLSIs (Springer, 2009), Mobile 3D Graphics SoC from Algorithm to Chip (Wiley, 2010), Bio-Medical CMOS ICs (Springer, 2011), Embedded Systems (Wiley, 2012), and Ultra-Low-Power Short-Range Radios (Springer, 2015). Dr. Yoo received the Electronic Industrial Association of Korea Award for his contribution to DRAM technology in 1994, Hynix Development Award in 1995, the Korea Semiconductor Industry Association Award in 2002, Best Research of KAIST Award in 2007, Scientist/Engineer of this month Award from Ministry of Education, Science and Technology of Korea in 2010, Best Scholarship Awards of KAIST in 2011, and Order of Service Merit from Ministry of Public Administration and Security of Korea in 2011 and has been co-recipients of ASP-DAC Design Award 2001, Outstanding Design Awards of 2005, 2006, 2007, 2010, 2011, 2014 A-SSCC, Student Design Contest Award of 2007, 2008, 2010, 2011 DAC/ISSCC. He has served as a member of the executive committee of ISSCC, Symposium on VLSI, and A-SSCC and the TPC chair of the A-SSCC 2008 and ISWC 2010, IEEE Fellow, IEEE Distinguished Lecturer ( 10-11), Far East Chair of ISSCC ( 11-12), Technology Direction Sub- Committee Chair of ISSCC ( 13), TPC Vice Chair of ISSCC ( 14), and TPC Chair of ISSCC ( 15).