검색 상세

A Interference-Tolerant Bio-potential Monitoring Analog Front-End

  • 모영진 (서강대학교 일반대학원)

초록(요약문)

Biopotential recording systems have become increasingly important in modern healthcare as they enable continuous and noninvasive monitoring of physiological signals such as electroencephalograms (EEG), electrocardiograms (ECG), electrooculograms (EOG), and electromyograms (EMG). Recent advances in wearable devices—including smartwatches, earbuds, and smart glasses—have further amplified the need for compact, low-power biopotential sensing circuits capable of operating reliably in real-life environments. Most wearable platforms rely on passive electrodes such as dry or non-contact electrodes, which exhibit large impedance variation and strong sensitivity to environmental interference. As a result, the analog front-end (AFE) that interfaces directly with these electrodes becomes the key block that determines overall system performance. To ensure robust signal acquisition, the AFE must maintain low noise, high input impedance, wide dynamic range, and low power consumption while rejecting large common-mode interference (CMI) and motion-induced artifacts (MA). However, achieving these requirements simultaneously is fundamentally challenging, and conventional AFE architectures often suffer from trade-offs among noise efficiency, input impedance, and common-mode robustness. Wearable biopotential monitoring systems require AFEs that operate reliably in the presence of large CMI and MA. In two-electrode measurement systems, where the bias electrode is removed to improve user comfort, the AFE becomes directly exposed to power-line CMI exceeding 100-VPP, and MA generated by electrode–skin interface variations often reaches tens of millivolts. These disturbances shift the amplifier operating point, reduce the effective dynamic range, and frequently cause output saturation. To address these challenges, this thesis presents two complementary AFE architectures that enhance interference robustness, suppress MA-induced distortion, and maintain linear operation under the stringent power constraints of wearable devices. The first contribution is a 4.6-µW biopotential AFE that achieves 133-VPP CMI tolerance by isolating the chip ground from earth ground and introducing a CMI-Follower providing a low common-mode input impedance path. Combined with the parasitic capacitance CGND, the CMI-Follower establishes a voltage-division mechanism that couples nearly the entire CMI to Chip-GND, theoretically minimizing the input-referred CMI without sacrificing noise performance. This architecture enables robust biopotential acquisition even under severe real-environment CMI. Although highly tolerant to CMI, this architecture remains susceptible to motion-induced baseline shifts and to residual CMI generated by variations in CGND. In wearable environments, electrode motion alters the electrode–skin impedance, introducing low-frequency baseline drift, while changes in the parasitic CGND modulate the voltage- division ratio and allow a portion of the CMI to reappear at the AFE input. To address these limitations, the second contribution introduces a MA tolerant AFE that prevents amplifier saturation and extends the ICMR. A low-power adaptive DC stabilization loop dynamically adjusts the input DC level based on output amplitude, ensuring a stable operating point in the presence of MA. In addition, an improved operational transconductance amplifier (OTA) structure expands the ICMR by 60 mV (12%) under a 1-V supply while maintaining noise efficiency. The digital feedback- loop state is also provided externally, enabling recovery of the original input during post-processing despite loop-induced distortion. Overall, the proposed AFEs significantly improve CMI tolerance, MA robustness, input dynamic range, and power efficiency, enabling stable long-term biopotential monitoring in real-world wearable applications. These techniques provide practical circuit-level design guidelines for next-generation low- power biomedical interfaces. Keywords: Bio-potential monitoring, analog front-end (AFE), chopper amplifier, electrode dc offset (EDO), DC servo loop (DSL), positive feedback, noise efficiency factor, two-electrode, common mode interference (CMI) tolerance, input dynamic range, total common mode rejection ratio (T-CMRR) Motion Artifact (MA).

more

목차

ABSTRACT １
List of Figures ５
List of Tables 7
List of Abbreviations 8
Chapter 1 : Introduction 10
Chpater 2 : Bio-potential Monitoring System １2
Chapter 2.1: Bio-potential Monitoring on Wearable Devices １2
Chapter 2.2: Bio-potential Characteristics １4
Chapter 2.3: Electrode Sensors for Bio-potential Monitoring １5
Chpater 3 : Capacitively-coupled Chopper Instrumentation Amplifier (CCIA) １7
Chapter 3.1: Basic Working Principles and Design Challenges １7
3.1.1 Basic Topology １7
3.1.2 Design Challenges １9
3.1.2.1 Electrode DC Offset (EDO) １9
3.1.2.2 Input Impedance ２1
Chapter 3.2: CCIA Design ２3
3.2.1 Analog DC Servo Loop (ADSL) ２3
3.2.2 Hybrid DC Servo Loop (HDSL) ２5
3.2.3 Positive Feedback Loop (PFL) ２7
Chpater 4 : CCIA Design with CMI and MA Suppression 30
Chapter 4.1: Introduction: Common Mode Interference (CMI) and Motion Artifact (MA) 30
4.1.1 Three Electrodes Model 30
4.1.2 Two Electrodes Model 32
4.1.3 Total Common Mode Rejection Ratio (T-CMRR) ３4
4.1.4 Motion Artifact (MA) ３6
4.1.5 Challenges in Wearable Systems ３7
Chapter 4.2: Previous Works for CMI Rejection and MA Suppression ３8
4.2.1 Pseudo Right-Leg-Driven (RLD) ３8
4.2.2 CMI Follower 40
Chapter 4.3: Proposed CCIA with CMA-CR-OTA and CF DDSL ４3
4.3.1 Operational Principles of proposed CMA-CR-OTA ４3
4.3.1.1 Concepet of proposed CMA-CR-OTA ４3
4.3.1.2 Circuit implementation of proposed CMA-CR-OTA ４3
4.3.2 Operational principles of proposed coarse fine digital DC servo loop (CF DDSL) ４6
4.3.2.1 Concepet of proposed CF DDSL ４6
4.3.2.2 Circuit implementation of proposed CF DDSL ４7
4.3.3 Circuit implementation of overall architecture ４9
Chapter 4.4: Experimental Results 50
4.4.1 Measurement Setup 50
4.4.2 Electrical Measurement Results of Proposed AFE ５2
4.4.3 Measurement Results of CMA-CR-OTA ５5
4.4.4 Measurement Results of CF DDSL ５6
4.4.5 Biopotential Measurement Results ５8
4.4.6 System Comparisons. 60
Chpater 5 : Conclusion and Limitations ６1
References ６2
Acknowledgements ６5

more
반출 Meta View 목록