超低功耗生物信号采集系统：模拟与算法融合设计

需积分: 9 121 浏览量更新于2024-07-17 收藏 6.77MB PDF 举报

身份认证购VIP最低享 7 折!

30元优惠券

《2019年模拟与算法辅助超低功耗生物信号采集系统》是一本深入探讨了利用模拟辅助和算法策略设计和实现超低功耗生物信号采集平台的专业著作。本书由Venkata Rajesh Pamula、Chris Van Hoof和Marian Verhelst共同编撰，是Analog Circuits and Signal Processing系列的一部分，该系列专注于出版关于模拟集成电路和信号处理电路及系统设计与应用的研究成果。书中特别关注的主题是"模拟与算法辅助"（Analog-and-Algorithm-Assisted）信号处理技术，这种方法通过在信号处理链早期采用低功耗生物信号采集系统和模拟预处理，结合算法策略，显著减少了能耗。这对于可穿戴传感器网络（WSN）和体域网络（BAN）尤其重要，它在降低传感器刺激功率消耗、数字信号处理以及无线通信链路的功耗方面展现出了明显的优势。书中的两个具体案例研究进一步展示了这种技术的应用：一是自适应采样心电图（ECG）采集系统，另一个是基于压缩采样的光体积描记术（PPG）采集系统。此外，书中还详细讨论了"超越奈奎斯特率"的信号采集和处理趋势，包括自适应采样和压缩采样等新理念。压缩域特征提取章节涵盖了光体积描记术作为一种新兴光学传感方式的处理，特别是在压缩采样PPG读取中嵌入特征提取的技术。此外，书中还探讨了传感器融合技术如何提升信号完整性，并降低生物信号采集系统的功耗。这反映了当前生物信号处理领域的前沿进展，旨在通过集成创新技术来推动行业的持续发展和节能目标的实现。《2019年模拟与算法辅助超低功耗生物信号采集系统》为读者提供了全面理解并实践这一重要领域的工具，对于电子工程师、研究人员、教育工作者和图书馆来说，是一本不可或缺的参考资料。随着对能源效率和无线健康监测设备的需求增长，这本书的研究成果对于推动绿色电子和可穿戴技术的发展具有重要意义。

资源详情

资源推荐

List of Figures

Fig. 1.1 Block diagram of a typical biosignal acquisition system and

its power breakdown ................................................. 2

Fig. 1.2 A standard ECG with its constituent entities labeled ............... 4

Fig. 1.3 PPG acquisition in: (a) transmission mode and (b)

reﬂectance mode ..................................................... 5

Fig. 1.4 Principle of transmission mode PPG acquisition ................... 5

Fig. 1.5 (a) Conventional biosignal acquisition and processing

paradigm. (b) Assisted processing architecture ..................... 7

Fig. 2.1 Concept of local bandwidth (BW) ................................... 12

Fig. 2.2 (a) Conventional biosignal acquisition system. (b) Biosignal

acquisition system employing adaptive sampling .................. 12

Fig. 2.3 Adaptive sampling applied to ECG signals. (Top)

Conventional uniform sampled ECG segment. (Bottom)

Adaptively sampled ECG segment .................................. 14

Fig. 2.4 Sampling and the uncertainty in peak position ...................... 15

Fig. 2.5 Signal quality and compression trade-offs between different

compression approaches for ECG signal. Adaptively

sampled ECG signal is reconstructed using ZOH .................. 17

Fig. 2.6 Digital implementation of the adaptive sampling controller ....... 17

Fig. 2.7 ECG acquisition system using adaptive sampling. It uses a

differentiator based activity detector to detect the presence

of QRS complex in the ECG waveform and alter the

sampling rate. The delay element compensates for the delays

in detection and alteration of sampling rate and enables

high-quality acquisition of QRS complexes ........................ 18

Fig. 2.8 First-order differentiator based activity detector C

=1pF

and C

= 100 fF ...................................................... 20

Fig. 2.9 Group delay dependence on ﬁlter order for Bessel ﬁlter ........... 21

Fig. 2.10 Fifth-order continuous time LC ladder structure with

double-termination Bessel ﬁlter (R

) ......................... 22

xvii

xviii List of Figures

Fig. 2.11 g

-C implementation of ﬁfth-order Bessel ﬁlter used as

delay element. PMOS differential pair is used to implement

the g

stages. All transconductances are 10.8 nS and C

20.08 pF C

=7.15pFC

=1.55pFC

= 9.89 pF and C

=4.51pF.............................................................. 23

Fig. 2.12 Small-signal equivalent circuit of an ideal OTA and

ﬁrst-order model with parasitics included .......................... 25

Fig. 2.13 Effect of parasitic poles on the frequency response of the

ﬁlter ................................................................... 26

Fig. 2.14 Microphotograph of the fabricated adaptive sampling

controller ............................................................. 27

Fig. 2.15 Measured transfer function of the delay element (g

-C ﬁlter) ..... 28

Fig. 2.16 Measured output voltage noise power spectral density of the

delay element (g

-C ﬁlter) .......................................... 28

Fig. 2.17 Total harmonic distortion of the delay element (g

-C ﬁlter) ...... 29

Fig. 2.18 Measured magnitude response of the differentiator in

activity detector ...................................................... 29

Fig. 2.19 A 5-s ECG signal from the record 101 from MIT-BIH

database captured by the ASC, reconstructed with ZOH. (a)

Measured output of the ASC with thresholds set to increase

the sampling rate in the presence of QRS complexes. (b)

Original and reconstructed signal (CR = 6, PRD = 0.12%).

(c, d) Measured output of the ASC in the presence of motion

artifacts with SNR = 6 dB and the reconstructed signal (CR

=6.5,PRD=0.15%).(e, f) Measured output of the ASC

in the presence of motion artifacts with SNR = 6 dB with

reduced threshold voltages and the reconstructed signal (CR

=4.1,PRD=0.12%)................................................. 30

Fig. 3.1 Uniform sampling of a continuous time signal ..................... 34

Fig. 3.2 Level-crossing based sampling of a signal .......................... 35

Fig. 3.3 (Top) Time domain PPG signal and (Bottom) its projection

on the DCT basis ..................................................... 37

Fig. 3.4 Signal acquisition in compressive sampling (CS) framework ..... 37

Fig. 3.5 Structured random matrices (Left) Bernoulli matrix (Right)

Binary matrix ......................................................... 38

Fig. 3.6 Typical use case scenario of CS in context of BAN ................ 41

Fig. 3.7 Proposed overlapped window reconstruction process .............. 44

Fig. 3.8 Original signal and the signals reconstructed using the

traditional approach (SNDR = 18.6 dB) and the proposed

approach (SNDR = 19.73 dB) ....................................... 47

Fig. 3.9 Average CPU execution time for traditional approach and

the proposed overlapped window approach over 25 runs

across 20 records ..................................................... 47

List of Figures xix

Fig. 3.10 Average SNDR for traditional approach and the proposed

overlapped window approach over 25 runs across 20 records ..... 48

Fig. 3.11 Analog CS implementation through RMPI ......................... 49

Fig. 3.12 Digital CS implementation .......................................... 50

Fig. 3.13 CS implementation through NUS ................................... 51

Fig. 3.14 Analog CS for EEG acquisition, exploiting spatial

correlation ............................................................ 52

Fig. 4.1 Various possible CS based acquisition systems in context of

BAN. (a) Signal acquisition is performed at the sensor node

while reconstruction and feature extraction is performed at

the base station. (b) Both CS encoding and decoding are

performed on the signal node followed by feature extraction.

directly from the CS data ............................................ 56

Fig. 4.2 Estimation of HR and HRV from time domain PPG signal ........ 59

Fig. 4.3 Estimation of average HR from the frequency spectrum of

PPG signal ............................................................ 59

Fig. 4.4 Average error in the inner product of randomly sampled

PPG signal normalized to the corresponding Nyquist rate

inner products ........................................................ 60

Fig. 4.5 Concept of frequency spectrum estimation from randomly

sampled signal using least-squares spectral ﬁtting ................. 61

Fig. 4.6 (a) Four time domain segments of PPG signal each 4 s

long. (b) Estimated spectrum from 10× randomly sampled

signals. α indicates the signiﬁcance level of the peak in

spectrum .............................................................. 63

Fig. 4.7 HRV using proposed approach and conventional approach.

HRV computed using proposed technique is highly

correlated to the one computed using conventional time

domain approach ..................................................... 64

Fig. 4.8 FFT of time domain PPG signals in Fig. 4.6a after

downsampling the signal by (a)10× (b)30×.DC

component is removed for clear representation ..................... 65

Fig. 4.9 Alias creation by downsampling the signal and the absence

of the same when random sub-sampling is employed .............. 66

Fig. 4.10 (a) Time domain PPG signal randomly sampled to obtain

30× compression. Sampling instants are highlighted. (b)

Estimated spectrum of the randomly sub-sampled signal .......... 66

Fig. 5.1 Conventional PPG acquisition system employing uniform

LED stimulation and sampling ...................................... 70

Fig. 5.2 CS based PPG acquisition system employing random LED

stimulation and sampling ............................................ 71

Fig. 5.3 Partial measurement matrix structure for CS PPG acquisition ..... 72

xx List of Figures

Fig. 5.4 LED and sampling pulse structure for conventional (uniform

sampling) and CS PPG acquisition .................................. 73

Fig. 5.5 The architecture of a single channel CS PPG acquisition

ASIC which embeds a DBE for feature extraction ................. 74

Fig. 5.6 Ideal PD and TIA interface. The TIA is formed by

employing resistive feedback around an OTA ...................... 75

Fig. 5.7 (a) Ideal PD and (b) a practical PD with shunt resistance

) and reverse bias junction capacitance (C

) ................... 75

Fig. 5.8 (a) Equivalent circuit for loop gain computation of TIA. (b)

Equivalent circuit for return ratio computation ..................... 76

Fig. 5.9 Magnitude response of A

and the reciprocal of feedback

fraction





.......................................................... 77

Fig. 5.10 Lead-lag compensated TIA. Stability margin is enhanced by

inserting a feedback capacitor (C

) into the network, which

introduces a left half plane (LHP) zero ............................. 77

Fig. 5.11 (a) Equivalent circuit for loop gain computation of the

lead-lag compensated TIA. (b) Equivalent circuit for return

ratio computation .................................................... 78

Fig. 5.12 Measured reverse bias junction capacitance of the

photodiode used in the current work as a function of reverse

bias voltage ........................................................... 79

Fig. 5.13 TIA interfaced with a current DAC (IDAC) at the input.

IDAC enables active rejection of I

, relaxing the channel

DR requirements ..................................................... 80

Fig. 5.14 Schematic of the 5-bit current DAC (IDAC) used to subtract

static component of photocurrent ................................... 80

Fig. 5.15 Two stage Miller compensated OTA with resistive source

degenerated current source in the ﬁrst stage. This OTA core

is used to realize the TIA and SI .................................... 81

Fig. 5.16 Switched integrator (SI) implementation in the current

work. The output of the TIA is further ampliﬁed and ﬁltered

through SI ............................................................ 82

Fig. 5.17 Switched integrator (SI) as noise limiting ﬁlter in the PPG

readout channel ....................................................... 83

Fig. 5.18 Schematic of the 12-bit SAR ADC with a unit capacitor of

800 fF ................................................................. 84

Fig. 5.19 The simpliﬁed architecture of the digital back end (DBE) ......... 85

Fig. 5.20 Key timing signals for AFE and LED driver control ............... 86

Fig. 5.21 Eight-way multiply accumulate (MAC) unit for accelerating

PSD estimation ....................................................... 87

Fig. 5.22 Chip micrograph of the CS PPG ASIC ............................. 88

Fig. 5.23 (Top) Measured output of the channel for a DC current

excitation of LED in uniform sampling mode. (Bottom)

Zoomed in view of response during one sampling instant ......... 89

剩余129页未读，继续阅读

drjiachen

粉丝: 171
资源: 2138

超低功耗生物信号采集系统：模拟与算法融合设计

Algorithm-Data-Structure-And-Algorithm.zip

The efficient-parallel stripe noise removal algorithm with low resource

facility-location-in-genetic-algorithm-master

facility-location-in-genetic-algorithm this project use genetic algorithm to

off-the-shelf exposure-fusion algorithm是

分别详细介绍以下的GCC编译选项的功能原理，每条不少于50字： -fira-algorithm=CB -fira-region=one -flifetime-dse=2 -freorder-blocks-algorithm=stc -fstack-reuse=none

62663-bilevel-optimization-algorithm

item-based collaborative filtering recommendation algorithm combining item c

program a 32-bit integer divide rapid algorithm without any float type in c langugage

解释minorization-maximization (MM) algorithm，并给出matlab代码编写的例子

[1] R. Zhang, L L. Song, L.J. Ran, Ultra-wideband Location Algorithm Based on Particle Swarm Optimization in Non-line-of-sight Environment [J].Sensors and Microsystems, 2017, 36(9): 117-120, 124.请把条参考文献的格式改成上面的那种

pentaho-aggdesigner-algorithm-5.1.5-jhyde.jar下载

jssp-genetic-algorithm

Which Linux functions are included in the schedule() scheduling algorithm

加密方式3DES,DES,AES,DSA,MD5,MQV,ECDSA,RSA,SHA-1,SHA-2,SHA-3,SHA-224,SHA-384,SHA-512,SHA-256,SM2,SM3,SM4,TLS分别是什么

docker cp cable-room-algorithm/ 0c:/home/

找几篇PSO优化k_means相关的论文

最新资源