Noncontact Detection of Vital Signs

and Vibrations Using Micro-Radar

Prof. Jenshan Lin

University of Florida

Gainesville, Florida

• Noncontact Detection of Vital Signs (Respiration and

Heartbeat)

– Detection Method

– History

– Examples

• Noncontact Measurement of Vibrations

– Nonlinear Doppler Phase Demodulation

– Experimental Verification

• Integration of Micro-Radar Sensor Chips

– PCB Modules

– CMOS Chips

– SiP with Antennas

• Applications

Outline

2

3

Video

This video explains:

http://news.ufl.edu/2008/12/03/baby-vital-signs/

How does it work? Doppler Effect!

Constant Velocity

Frequency Shift

Periodic Chest Wall Movement

Phase Shift

( )R t

( )T t

0d( )x t

0( )x t v t

4

( ) sin( )x t t

x(t)

T t ft t ( ) cos 2 ( )

d dx tR t ft t

c

0 04 24 ( )

( ) cos 2 ( )

d

0

4 x t

4 ( )Antenna-to-target

round-trip delay Phase modulation due to

chest-wall movement

• Mix received signal with part of transmitted signal as

reference in mixer (similar to direct down-conversion):

A Simple Detection Method

T t ft t ( ) cos 2 ( )

d dx tR t ft t

c

0 04 24 ( )

( ) cos 2 ( )

T(t) x R(t) Baseband Signal B(t):

d x tB t

00

4 4 ( )( ) cos

0 d

t tc

02

( ) ( )Phase delay in receiver circuit

5

T(t)

R(t) B(t)

Small Angle Approximation (Linear)

4 ( ) 4 ( )( ) sin

x t x tB t

It is similar to measuring phase noise using

FM discriminator technique.

6

0

04)(4cos)(

dtxtB

90, 270 , … When …

Small

x(t) <<

negligible

• Fortunately, if the distance between radar sensor and

target is small enough, close-in phase noise (slow

variation) of R(t) and T(t) are correlated.

Range Correlation Effect

How about ?

T t ft t ( ) cos 2 ( )

d dx tR t ft t

c

0 04 24 ( )

( ) cos 2 ( )

d x tB t

00

4 4 ( )( ) cos ~ 0 @ short-range

A. D. Droitcour, O. Boric-Lubecke, V. Lubecke, J. Lin, G. Kovacs, “Range Correlation and I/Q

performance benefits in single chip silicon Doppler radars for non-contact cardiopulmonary signs

sensing,” IEEE Trans. Microwave Theory Tech., Vol. 52, No. 3, pp. 838-848, March 2004 7

Earliest Research Report

8

• J. C. Lin, "Noninvasive Microwave Measurement of Respiration," Proceedings of the IEEE, vol. 63, no. 10, p. 1530, Oct. 1975.

Rabbit

Human

Early Research Effort

9

• K.-M. Chen and H.-R. Chuang, “Measurement of Heart and Breathing Signals of Human Subjects Through Barriers with Microwave Life-Detection Systems,” IEEE EMBC 1988. – 10GHz: 1.5 ft of dry bricks – 2GHz: 3 ft of dry bricks

• H.-R. Chuang, Y.-F. Chen, and K.-M. Chen, “Automatic Clutter-Canceler for Microwave Life-Detection System,” IEEE Trans. Instrumentation and Measurement, Vol. 40, No. 4, August 1991.

10 GHz

RF Out to

Tx Antenna

Baseband +

Baseband -

RF In from

Rx Antenna

Buffer LO

Buffer

VCORF

Buffer

Mixer

RF Out to

Tx Antenna

Baseband +

Baseband -

RF In from

Rx Antenna

Buffer LO

Buffer

VCORF

Buffer

Mixer

• 0.25 µm BiCMOS

• 3.75mm x 3.75mm

• 1.6 GHz transmitted

• Output power = 6.5dBm

First Non-Contact Vital Sign Sensor Chip

10

Balun

Balun

Balun

VCO

Mixer

RF out

(to Tx antenna)

RF in

(from Rx antenna)

Baseband out

Test Structures

VCO

LO Buffer

Buffer

RF Buffer

Mixer

Test Structures

VCO

LO Buffer

Buffer

RF Buffer

Mixer

A. D. Droitcour, O. Boric-Lubecke, V. Lubecke, J. Lin, “0.25m CMOS and BiCMOS Single Chip

Direct Conversion Doppler Radars for Remote Sensing of Vital Signs,” IEEE International Solid

State Circuits Conference Digest of Technical Papers, pp. 348-349, 2002.

Free-running VCO. No PLL.

• 0.25 µm CMOS

• 2.4 GHz transmission frequency

• Direct-conversion – no IF and no image-reject filter

• Free running VCO – no PLL, no crystal.

• Quadrature receiver – to avoid null-point problem.

2nd Non-Contact Vital Sign Sensor Chip

11

0

1

2

3

4

5

0 2 4 6 8Time (sec)

Am

plit

ude

A. Droitcour, O. Boric-Lubecke, V. Lubecke, J. Lin, G. Kovacs, “Range Correlation Effect on ISM Band I/Q

CMOS Radar for Non-Contact Vital Signs Sensing,” IEEE MTT-S International Microwave Symposium

Digest, Vol. 3, pp. 1945-1948, 2003.

2003 – present

Research @ University of Florida

12

CW Radar Carrier Frequency

4 ( ) 4 ( )4 ( ) 4 ( )( ) cos( )h hr r

x t x tx t x tB t

Y. Xiao, J. Lin, O. Boric-Lubecke, V. Lubecke, “A Ka-Band Low Power Doppler Radar System for

Remote Detection of Cardiopulmonary Motion,” Proceedings of the 27th IEEE Engineering in Medicine

and Biology Society Annual International Conference, pp. 7151-7154, 2005. 13

• Short wavelength is more sensitive to small displacement

- improve signal level while keeping transmitted power low.

Increase the carrier frequency.

Long wavelength

Small

Short wavelength

Large

Same x

x x

Double-Sideband Transmission/Detection

Example @ Ka-Band

Baseband Output

B(t)

LO2 LO1

LNA Mixer IF_AMP

Mixer

Mixer AMP BPF AMP

(0.1-10 Hz) (26-40 GHz)

Double-sideband transmission and detection

no image rejection needed simple architecture

Feasible for monolithic integration on one chip

RF

0 5 10 15 20 25 -3

-2

-1

0

1

2

3

Time (Sec)

Dete

cte

d S

ignal (V

) LO2+LO1

LO2-LO1

14

Ka-band Bench-Top System

LabVIEW

Ka radar

Antenna

Subject

under test

Baseband

DAQ

15

User Interface Labview – data acquisition and signal processing

16

Measurement from Four Sides

Antenna

BackFront

Left

Right

LSB: 26.54 GHz

USB: 27.66 GHz

• Two power levels: 14.2 μW, 350 μW

• Five distances: 0.5m, 1m, 1.5m, 2m, 2.5m

Y. Xiao, C. Li, J. Lin, “Accuracy of A Low-Power Ka-Band Non-Contact Heartbeat Detector

Measured from Four Sides of A Human Body,” IEEE MTT-S International Microwave

Symposium Digest, pp. 1576-1579, June 2006. 17

Typical Test Result

0 5 10 15 20 25-3

-2

-1

0

1

2

3

Time (Sec)D

ete

cte

d S

ign

al (V

)

0 5 10 15 20 2580

85

90

95

Time (Sec)

Be

ats

/Min

Detected heart beatReference heart beat2 % Higher than

Reference

2 % Lower thanReference

Time-Domain Signal

Calculated Heart Rate

18

Test Results

SUMMARY OF HEART RATE DETECTION ACCURACY Distance (m) Front Left Right Back

Double sideband transmitting power: 14.2 µW

0.5 99.1% 96.3% 100% 97.6%

1 89.8% 89.8% 93.2% 100%

1.5 98.9% 89% 93.8% 94.3%

2 85.2% 80.5% 97.4% 93.6%

2.5 83.3% 85.7% 85.1% 85.5%

Double sideband transmitting power: 350 µW

0.5 100% 100% 100% 100%

1 94.8% 94.7% 93.2% 100%

1.5 98.1% 97.6% 100% 100%

2 100% 100% 100% 100%

2.5 95.1% 100% 95.2% 97.2%

19

• Very good accuracy achieved with very low transmission power.

• Accuracy better than 80% from any side, at any distance, and under

either power level.

• Measurement from the back: the best performance! WHY?

Interesting Observations

0 20 40 60 80 100 1200

0.2

0.4

0.6

0.8

1

Beats/Min

No

rma

lize

d S

pe

ctr

um

Breathing Fundamental

Breathing 2nd Harmonic

Breathing 3rd Harmonic

Heartbeat

0 20 40 60 80 100 1200

0.2

0.4

0.6

0.8

1

Beats/Min

Breathing Fundamental

Breathing 2nd Harmonic

Breathing 3rd Harmonic

Heartbeat

Front Back 20

Nonlinear Doppler Phase Demodulation

and

Measurement of Vibrations

21

Nonlinear Doppler Phase Demodulation

when = 90 and

xh(t), xr(t) <<

xr(t) <<

However, at high frequency (short wavelength),

the displacement might not be small enough and

the small angle approximation might not be valid.

4 ( ) 4 ( )( ) cos( )

4 ( ) 4 ( )

h r

h r

x t x tB t

x t x t

22

Small angle approximation:

Nonlinear Doppler Phase Demodulation

4 ( ) 4 ( )( ) cos( )

4 4( ) ( )cos( )

h r

h rl k r h

k l

x t x tB t

m mJ J k t l t

Periodic body movements due to heartbeat and respiration

xh(t) = mh·sinωht, xr(t) = mr·sinωrt

f

k=1

l=0k=2

l=0k=3

l=0

k=0

l=1

0

k=0

l=0

k=-1

l=1

k=1

l=1

R1 R3 H1 R2 C1a C1b

C. Li, Y. Xiao, J. Lin, "Experiment and Spectral Analysis of a Low-Power Ka-Band Heartbeat Detector

Measuring from Four Sides of a Human Body," IEEE Transactions on Microwave Theory and

Techniques, IMS2006 Special Issue, Vol. 54, No. 12, pp. 4464-4471, December 2006. 23

Accurate Measurement of Periodic Motion

• For a single tone vibration, magnitudes of harmonics

in baseband spectrum are determined by:

– Residual phase

– Wavelength

– Displacement of vibration

1 2 3 4 1 2

3 4

4 4: : : | ( )cos |:| ( )sin |:

4 4| ( )cos |:| ( )sin | .

m mH H H H J J

m mJ J

The displacement of vibration can be accurately

determined from the ratio of harmonics!

C. Li, J. Lin, "Non-Contact Measurement of Periodic Movements by a 22-40GHz Radar Sensor Using

Nonlinear Phase Modulation," IEEE MTT-S International Microwave Symposium Digest, pp. 579-582,

June 2007 24

H1 H2 H3 H4

Measurement Example

0 5 10 15-0.5

0

0.5

1

Time (Second)

Ba

se

ba

nd

Sig

na

l (V

)

0 1/3 2/3 1 4/3 5/3 20

0.5

1

Frequency (Hz)

No

rma

lize

d S

pe

ctr

um

1 1.5 2 2.5 30

1

2

3

4

5

Displacement (mm)

Ha

rmo

nic

Ra

tio

H3/H1 ratioH4/H2 ratio

Meaured H3/H1

Meaured H4/H2

• Movement period T = 3 sec amplitude = 2 mm

• fRF: 40 GHz,

• Transmission power: 50 µW

• Distance: 1.65m

(a) Baseband signal, spectrum (b) Displacement extraction,

self-verification 25

2.056 mm

2.045 mm

• PCB modules

• IC chips: RFIC

• System-in-Package: Antennas integrated

Development of Integrated Radar Sensors

at University of Florida

26

PCB Modules

27

External ADC and Signal Processing:

Output connected to data acquisition

module (DAQ) and notebook computer.

Both radar module and DAQ can be

powered by USB cable, external power

supply/charger, or battery.

On-board ADC and Signal Processing:

Onboard ARM processor and ADC. No

external DAQ and notebook computer

needed. Powered by a battery.

Search and Rescue Robot

28

Wireless data link sends detected data to a remote station.

Double-Sideband Radar Sensor Chip

IF-Amp LNA

IF-VCO RF-VCO

Bas

eban

d

IF-Mixer

Test

Down-Conv

Up-Conv

G=18 dBG=23 dB

1f 2f

5 GHz Radar Chip

Balun

f1 = 60~520 MHz; f2 = 4.6~5.7 GHz UMC 0.18 µm CMOS

1.3×1.6 mm2

C. Li, Y. Xiao, J. Lin, "A 5-GHz Double-Sideband Radar Sensor Chip in 0.18-µm CMOS

for Non-contact Vital Sign Detection," accepted, IEEE Microwave and Wireless

Components Letters, 2008 29

Test Result – Output Spectrum

4 4.5 5 5.5 6 6.5-100

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

Freq. [GHz]

Am

p.

[dB

m]

Res BW 100 kHz VBW 100 kHz Sweep 301.4 ms(601 pts)

Atten10dB

Mkr1 5.467 GHz; Mkr2 5.683 GHz; Mkr3 5.900 GHz

-7.16 dBm; -30.51dBm; -6.14 dBm

4 4.5 5 5.5 6 6.5-100

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

Freq. [GHz]

Am

p.

[dB

m]

Res BW 100 kHz VBW 100 kHz Sweep 301.4 ms(601 pts)

Atten10dB

Mkr1 4.417 GHz; Mkr2 4.625 GHz; Mkr3 4.825 GHz -12.50 dBm; -29.95dBm; -12.40 dBm

RF VCO tuned to the lowest RF VCO tuned to the highest

30

Spurs will be out of band and filtered out after down-conversion

Test Result – Vital Sign Detection 0 3 6 9 12 15

-1

0

1

0 20 40 60 80 1000

0.5

1

Beat/Min

Sp

ectr

um

Respiration

Harmonic Heartbeat

0 3 6 9 12 15-1

0

1

0 20 40 60 80 1000

0.5

1

Beat/Min

Sp

ec

tru

m

Respiration

Harmonic Heartbeat

(a)

0 3 6 9 12 15-0.5

0

0.5

0 20 40 60 80 1000

0.5

1

Beat/Min

Sp

ec

tru

m

Respiration

Heartbeat Harmonic

0 3 6 9 12 15-0.5

0

0.5

0 20 40 60 80 1000

0.5

1

Beat/Min

Sp

ectr

um

Respiration

Heartbeat Harmonic

(b)

0.5 m

2 m

31

Radar Receiver with Gain Control

inRF

outI +

outI -

outQ +

outQ -

Load Clk DataQ

LO +Q

LO

ILO +

ILO

LNAPre-Amp

MixerI

MixerQ

VGA

I

VGA

Q

Gm-boosted

biasBandgap Const-Gm

3-wire control

Gm Boosted Bias

Bandgap Constant Gm

LNA PreAmp

Mixer VGA

3 Wire Control Unit

RFin

Clock Data Load

Q

I

DAQ CPLD

5.8 GHz Radar Receiver

UMC 0.13 µm CMOS

1.2×1.2 mm2

C. Li, X. Yu, D. Li, L. Ran, J. Lin, "Software Configurable 5.8 GHz Radar Sensor Receiver

Chip in 0.13 μm CMOS for Non-contact Vital Sign Detection," IEEE RFIC Symposium

Digest of Papers, June 2009

32

Test Result

Power: 1.5 V battery

3-wire program: Xilinx XC9536 CPLD

Antenna: 2-by-2 patch array, 9 dB gain

DAQ: NI USB-6008, 12 bit, 0-5V input

0 5 10 15 20 25

-20

0

20

Time (Second)

I/Q

Sig

na

l [m

V]

0 20 40 60 80 100 1200

0.5

1

Beats/Min

CS

D S

pe

ctr

um

Respiration Heartbeat

0 5 10 15 20 25-40

-20

0

20

Time (Second)

I/Q

Sig

na

l [m

V]

0 20 40 60 80 100 1200

0.5

1

Beats/Min

CS

D S

pe

ctr

um

Respiration

Heartbeat

Detect from the back @ 0.5m away Detect from the front @ 1.5m away 33

Flip-chip radar transceiver (UMC 90nm CMOS), two PCB patch

antennas, and DC biasing through blue wires.

Weight less than 10 gram (0.3 ounce). Size 31.2 mm x 45 mm.

60 GHz Flip-Chip Integrated Micro-Radar

34

T.-Y. J. Kao, A. Y.-K. Chen, T.-M. Shen, Y. Yan, J. Lin, "A Flip-Chip-Packaged and Fully

Integrated 60 GHz CMOS Micro-Radar Sensor for Heartbeat and Mechanical Vibration

Detections," IEEE RFIC Symposium Digest of Papers, June 2012.

Block Diagram of 60 GHz Micro-Radar

54 GHz

RF VCO

LNA

Balun

Driver On-chip

60 GHz RF

Passive

Mixer

6 GHz IF

QVCO

LO_I

I(t)

Q(t)

S

G

S

G

GG

G

S

S

G150 μm

150

μm

Top View (Not to scale)

Zoom-in

View

RT/duroid 5870 Laminate

GND

GND

GND

1.66 mm

1.8 mm

λ/4

TX

RX

Flip-chip

AttachPatch

Antenna

6 GHz IF Baseband

RF Pads

S

G

S

G

G

175 μm

31.3 mm

50 Ω

Microstrip

45

mm

0.9

6 m

m

2.35 mm

Active

Mixer

LO_Q

LO_I

Indirect down-conversion with passive mixers – low flicker noise

I/Q generation at IF

Same VCOs for up- and down-conversion - range correlation effect

Experiment – I/Q Channel Test

Actuator vibrating at distance (D) = 0.3 m away

displacement (A ) = 1 mm, f = 1 Hz

Q near optimal point,

I near null point

I near optimal point,

Q near null point

36

Complex Signal Demodulation

Optimal and null detection points

Complex signal demodulation (CSD):

( ) ( ) ( )

4 ( ) 4 ( )cos + sin

4 ( ) 4 ( )= exp = exp .exp

t t

t t

S t I t j Q t

x t x tj

x t x tj jj

Quadrature(I/Q) radar receiver needed

4 ( )( ) cos .t

x tB t

Combining I and Q channels

37

Experiment – Small Mechanical Vibration Detection

38

Vibration (f = 1 Hz, D = 0.3m):

→ minimum “detectable”

displacement A = 20 μm

All data points are

normalized to the largest

CSD spectrum peak (A = 1

mm, D = 0.3 m)

→ A = 0.2 mm can be

detected at D =2.1 m away

increase

distance

Experiment – Heartbeat Detection

Subject sitting on a chair 0.3 m in

front the radar.

At t = 0 ~ 7 sec, Q was around the

optimal point, and I was near the

null point.

After t = 9 sec, I channel started

to take over the detection due to

slight body movement (null point

every λ/4).

Baseband noise voltage is around

1mVrms, corresponding to a

baseband SNR around 5 (14 dB)

(Holding breath)

0 2 4 6 8 10 12 140.19

0.2

0.21

0.22

Q (

v)

Time (s)

0 2 4 6 8 10 12 140.54

0.55

0.56

0.57

I (v

)

Time (s)

0 20 40 60 80 100 1200

0.5

1

CS

D S

pectr

um

Beat / Minute

Heartbeat Body

Movement

39

Problem of Random Body Movement

During Vital Sign Measurement

A Solution…

40

Random Body Movement Cancellation – Concept

Heart

fI fQ

090

bIbQ

DAQ DAQDemodulation

090Body

41

• Cardiorespiratory movements on both sides of the body move in the

same direction w.r.t to their detecting radars

• Random body drift movements are in the opposite directions w.r.t. to

their detecting radars

Random Body Movement Cancellation – Theory

1 11

4 ( ) 4 ( ) 4 ( )( ) exp h r

f

x t x t y tS t j

2 22

4 ( ) 4 ( ) 4 ( )( ) exp h r

b

x t x t y tS t j

1 2 1 2

1 2

4 ( ) ( ) 4 ( ) ( )( ) ( ) ( ) exp h h r r

fb f b

x t x t x t x tS t S t S t j

Front:

Back:

Combine signals from both sides:

xh1, xh2, xr1, xr2: physiological movements

y(t): random body movement

1hx 2hx

1rx2rx

y

42

y(t) (random body movement) disappeared!

Random Body Movement Cancellation – Experiment

Body

ResHeart

fI fQ

090

bIbQ

DAQ DAQDemodulation

Res0

90

43 Isolation between two radars: antenna polarizations & slight frequency

offset between TX and RX

Random Body Movement Cancellation – Result

0 20 40 60 80 100 1200

0.5

1

Beats/Min

Co

mb

ine

d S

pe

ctr

um

Recovered heartbeat

(b)

0 20 40 60 80 100 1200

0.5

1

Beats/Min

Sp

ectr

um

FrontBack

(a)

0 2 4 6 8 10 12-1

-0.5

0

0.5

1

Time (Second)

Am

plit

ud

e

I ChannelQ Channel

(a)

0 2 4 6 8 10 12-1

-0.5

0

0.5

Time (Second)

Am

plit

ud

e

I ChannelQ Channel

(b)

Front

Back

C. Li, J. Lin, “Random Body Movement Cancellation in Doppler Radar Vital Sign

Detection,” IEEE Transactions on Microwave Theory and Techniques, vol. 56,

issue 12, pp. 3143-3152, December 2008. 44

First Demonstration of Noncontact Vital Sign

Measurement on Treadmill

45

τd

SIL Radar 1

Sbb1(t)

SIL Radar 2

Not

usedτd

d0 = 1m

s0 = 2m

SIL: Self-Injection Locking

F.-K. Wang, T.-S. Horng, K.-C. Peng, J.-K. Jau, J.-Y. Li, and C.-C. Chen, "Single-

Antenna Doppler Radars Using Self and Mutual Injection Locking for Vital Sign

Detection With Random Body Movement Cancellation," Microwave Theory and

Techniques, IEEE Transactions on, vol. 59, pp. 3577-3587, 2011.

Applications

• Search-and-rescue

• Human healthcare; Animal care

• Radar in cell phone - lie (or emotion) detection radar

• Sports, video games, …

• When the complete radar system can be made very small… – small radar sensor chip + small antenna + small robot + wireless

ad hoc network + wireless energy or energy harvesting

– Searching survivors under rubbles will be much more effective with a swarm of small robots, e.g., robotic ants.

• Doppler micro-radar sensors have been demonstrated – PCB modules, RFIC, System-in-Package (SiP).

• Small micro-radars can be added to many electronic devices – computers, phones, tablets

• The technology can be used to detect any motion of an object reflecting radio waves

• With proper signal processing, useful and interesting information can be extracted for various applications (biomedicine, biometrics, …).

• New hardware architectures and sign processing algorithms are being developed by many groups in the world.

Summary

47

48