Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
12/22/2011
1
Google, iPhoneWhat Comes Next?
Prof. Yong LIAN, IEEE Fellow, Provost’s Chair Professor,
Area Director, Integrated Circuits & Embedded Systems
Editor-in-Chief, IEEE Trans. on Circuits & Systems II
Founder, ClearBridge VitalSigns Pte Ltd
Quiz 1: What is zettabyte?
Answer: 1021 or 270
Name ValueKilobyte (kB) 103
Megabyte (MB) 106
Gigabyte (GB) 109
Terabyte (TB) 1012
Petabyte (PB) 1015
Exabyte (EB) 1018
Zettabyte (ZB) 1021
Yottabyte (YB) 1024
12/22/2011
2
Google to
acquire Motorola Mobility
for $12.5B
Apple sues
Samsung over
Galaxy products
Apple Market Value (2011)
$363 billion
Google Market Value (2011)
$158 billion
By the end of 2015, annual global IP traffic will reach the zettabyte threshold
Internet Traffic Trend
12/22/2011
3
Network Traffic in 2015
0% 20% 40% 60% 80%
Internet
Managed IP
Mobile dataBusiness
Consumer
Source: Cisco VNI, 2011
What Comes Next?
Source: http://www.youtube.com/watch?v=-2faggNVQtM
12/22/2011
4
Basic Needs
Social Needs Self-actualization
Self-esteem
Social
Maslow’s Pyramid in Internet Era
Physiological
Safety
Computer
Anonymity
Links/Followers
Creativity
Expression
Basic Needs
How about basic needs?-Your body
-Your wellness-Your health
Maslow’s Pyramid in Internet Era
Physiological
Safety
Computer
Anonymity
12/22/2011
5
• Chronic diseases constitute more than 60% of deaths world
Are the Basic Needs Met?
In Singapore, 2 in 5 adults aged > 20 are already suffering from at least one chronic ailment
Source: World Economic Forum and the Harvard School of Public Health, 2011
Productivity Loss: 47 trillions by 2030
Source: World Economic Forum and the Harvard School of Public Health, 2011
12/22/2011
6
Health Expenditure (% Gross Domestic Product)
Source: OECD Health Data 2011
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
World Population
Source: Science, 333, 569 (2011)
12/22/2011
7
The World Population Is Getting Older?
Source: UNFPA, State of World Population 2011
Working-Age Adults per Older Adults
Source: Science, 333, 569 (2011)
12/22/2011
8
Challenges
• Largest issues faced in society today– Growing number of people with chronic
conditions
– Increasing healthcare costs
– Ageing populations
• Healthcare systems are designed for “repair works”, not for maintenance
• Hospital services are in the form of “point of care”, no continuous effort
Unmet Needs
• Needs from healthcare provider– low cost, reduced workload, and efficient
services
• Needs from individual– Personalized and prevention-oriented
healthcare
− Be well, keep well, get well
– Continuous monitoring and instant feedback from healthcare provider
12/22/2011
9
The Impacts of Electronics
Source: New York Stock & Commodity Exchanges
100
1000
10000
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
Do
w J
on
es A
vera
ge
Silicon chip InternetTCP/IP
PC
Mobile phone iPhone
Developments in Biomedical Devices
• Pacemaker (1958)
• Cochlear implants (1984)
• Deep brain stimulation (1997)
• Retina implants (2007)
• e-nose
• Prosthetic arms
• Prosthetic knees
12/22/2011
10
Bioelectronics
Pioneer of brain stimulation: Dr. Jose Delgado
Source: http://www.youtube.com/watch?v=6nGAr2OkVqE
Bioelectronics
Deep brain stimulation for Parkinson’s Disease
Source: http://www.youtube.com/watch?v=uMaCiuapAW0
12/22/2011
11
The Next Wave: myHealth• Seamless integration with home, working, and hospital
environments.• Continuous monitoring and instant feedback
• Prevention-oriented healthcare
– Prevention is better than cure
• Personalized service
• Key to the success
– New healthcare business model
– Innovations in engineering
– Change of mindset
myHealth: A Melting Pot
12/22/2011
12
Wireless Biomedical Sensor System
Human Energy Scavenging
• Wearable devices can generate 0.3mW – 8 W from breathing, finger motion, blood pressure, body heat, walking.
• Implantable nanowire devices:– Current density: ~8.9nA/ cm2
– Output voltage: ~96mV– Power density: 2.7mW/ cm3
Z.L. Wang, et al “Self-powered nanowire devices”, Nature Nanotechnology, Mar 2010.
12/22/2011
13
Challenges in Designing of Self-Powered WBS
• Complicated system contains analog, mixed-signal, digital, RF, and power blocks.
• Energy scavenging from body and ambient are unstable.
• Limited power budget– Less than few mW for wearable devices.
– Less than 1 mW for implants
Possible Solutions
• New system architecture• Exploring new signal processing flow:
continuous-in-time and discrete-in-amplitude.– Event driven ADC with continuous-time digital
signal processing achieves up to 80% dynamic power saving in terms of number digital outputs for audio signals*.
• Cross-domain optimization and processing• Wireless communication through human body • Asynchronous architecture.• Sub-threshold circuits.*M. Kurchuk and Y Tsividis, “Signal-dependent variable-resolution clockless A/D conversion with application to continuous-time digital signal processing”, IEEE Trans. on CAS I, May 2010 .
12/22/2011
14
A Snapshot of Research in ECE
• 1-V 450nW programmable ECG sensor interface
• 1-V 2.3µW ECG-on-Chip
• 1-V 22µW 32-channel ECoG chip
• 0.5-V 18µW 16-channel neural recording chip
• Sub-mW UWB mostly digital wireless transceiver
• 0.5-V low power ADC (21 fJ/conversion-step)
• 300mV event driven ADC
• 250mV digital filter for bio-signal processing
• Computationally efficient DSP algorithm (QRS detection at 500 nW for ASIC implementation)
A Design Example: 0.5-V, 1.13-μW per Channel Multi-Channel Neural
Recording Chip
12/22/2011
15
How Many Neurons in the Brain?
~100,000
Source: http://www.wikipedia.org
~16,000,000~200,000,000,000 ~100,000,000,000
~11,000
Neural Recording
Utah array from Cyberkinetics Inc
Michigan array from NeuroNexus Inc
12/22/2011
16
• One channel, one ADC
Multi-Channel Acquisition Systems
An
alo
g MU
X
Analog Multiplexing
• Sharing one ADC among all channels –significant increase in power consumption due to additional buffers
12/22/2011
17
Digital M
UX
Proposed Digital Multiplexing• Independent S/H stage for every channel, while
sharing one large DAC in SAR ADC
Timing Diagram
Significant increase in sampling time for every channel power reduction
AD conversion for every channel in sequential order
12/22/2011
18
System Architecture
MU
X
8-channel building block to form a larger array of recording interface
Instrumentation Amp. & VGA
Large gate area and high biasing current to lower noise floor
VGA as driving buffer
12/22/2011
19
8-Channel 8-bit SAR ADC
SARC02C04C08C0128C0 113C0
8-bit DAC Array
rst rst S3 S2 S1 S0Dout
C02C04C08C0 C0
4-bit S/H Array
S7 S6 S5 S4
Bootstrapped
Ch[0]
VDD
VDD
Ch[1]
Ch[6]Ch[7]
S4~S7Latch
S0~S3rst
Sample[0]~Sample[7]
Sample[0]
Cp[0]
8
MUX
Chip Prototype
0.13µm CMOS → Core area: 1.17mm2
12/22/2011
20
Frequency Responses
101
102
103
104
30
40
50
60
Frequency (Hz)
Gai
n (d
B)
Gain: 48/54dB
LPF: 7.5kHzHPF: 40~400Hz
Input Referred Noise
101
102
103
104
10-8
10-7
10-6
Frequency (Hz)
Inpu
t R
efer
red
Noi
se
(
V/
Hz)
5.32µVrms (50~12kHz)
12/22/2011
21
DNL & INL
0 32 64 96 128 160 192 224 255-0.5
-0.250
0.250.5
INL
(LS
B )
Output Code
0 32 64 96 128 160 192 224 255-0.5
-0.250
0.250.5
Output Code
DN
L(L
SB
)
DNL/INL: 0.34/0.42 LSB
ESSCIRC 2011, Helsinki42
Output PSD
0 5 10 15-100
-75
-50
-25
0
Frequency (kHz)
Out
put
PS
D(d
BF
S)
fin=14.001kHz (full scale),
SNDR=45.8dB, SFDR=71.26dB,ENOB=7.32b
FOM: 21fJ/Conversion-step
12/22/2011
22
Multi-Channel Performance0
128255
0128255
0128255
0128255
Out
put C
ode
0128255
0128255
0128255
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
0128255
Time (s)
Extra-Cellular Recording
Single neuron recording
12/22/2011
23
Performance Comparison
Parameter Harrison’07 Chae’08 Lee’10 Muller’11 This work
Supply (V) 3.3 3.3 3 0.5 0.5
Process (µm) 0.5 0.35 0.5 0.065 0.13
FE Current (µA) 12.8 2 25 - 1.72
Input referred noise (µVrms)
5.1 4.9 4.39 4.9 5.32
NEF - - - 5.99 3.09
ADC res. & Samp. rate per
ch.
10b15kS/s
6~9b40kS/s
--
8b20kS/s
8b30kS/s
No. of channel 100 128 32 1 16
Area per channel (mm2)
0.15 0.36 0.26 0.013 0.073
Avg. power per channel (µW)
142.2 23.4 36.6 5.04 1.13
The Future of Bioelectronics
Fantastic Voyage (1966)Source: http://www.youtube.com/watch?v=3o8vsU0Dw-4
12/22/2011
24
Robot with Rat Neuron
Source: http://www.youtube.com/watch?v=1-0eZytv6Qk
The Future of Bioelectronics
Source: http://www.nanowerk.com/spotlight/spotid=15292.php
12/22/2011
25
The Future WBS
100µm 10 µm 1 µm 100 nm 10 nm 1 nm
Human hair
MOSFET
ProteinsBacteria
Cell Virus
DNA Atoms
The Future of Biomedical Circuits & Systems
• Closed-loop diagnostic and therapeutic devices
• New circuit techniques for low voltage operation and 3D IC integration
• Wireless communication via human body
• System-in-package for miniaturized ingestible/injectable devices
12/22/2011
26
Conclusions
• The next wave of technology may come from healthcare, especially on wireless health and biomedical circuits and systems.
• Expanding the frontiers of biomedical circuits and systems to ride on the new wave– Many challenges in system architecture, low voltage
circuit techniques.– Call for revolutionary signal processing flow that
improves energy efficiency.– Be aware of challenges beyond CMOS: system
integration, packaging, etc.– Multidisciplinary is the key to the success.
Acknowledgment
• Students, Research Engineers, Postdocs– Chunzhu Yang, Liang Dong, Jianghong Yu, Ling Cen, Yingping Hu,
Ying Wei, Xiaohua Tian, Xiaodan Zou, Jinghua Zhang, Wen Sin Liew, Jun Tan, Fei Zhang, Chacko John Deepu, Mahmood Khayatzadeh, Lei Wang, Yibin Hong, Yongfu Li, Yonghong Tao, Xiaoyang Zhang, Zhe Zhang, Zhenglin Yang, Lei Zhang, Jiqing Cui, Yunfeng Liang, Xiaoyun Liu, Xiangdong Zhou, Pinping Sun, Praveen Thomas, Chandrasekaran Rajasekaran, Rui Cao, Jun Gu, Jye Sheng Hong, Wei Seng Chew, Saravanan s/o Velayutham, Zhi Li, Murli Nair, Seyed Hossein Seyedmehdi, Xiaoyuan Xu, Chao Xue, Muhammad Cassim Mahmud Munshi, Qi Zhang, Xiang Cheng, Junle Pan, Ashton Wong, Xiaofei Chang, Ti Li, Daren Zhang.
• Sponsorship– National University of Singapore research fund
– Agency of Science, Technology and Research (ASTAR)