IMPROVING PERFORMANCE OF BIO-RADARS FOR REMOTE HEARTBEAT AND BREATHING DETECTION BY

USING CYCLOSTATIONARY FEATURES

Daniel Malafaia

José Vieira

Ana Tomé

CReATION - Cognitive Radio Transceiver Design for Energy Efficient Data Transmission

EXCL/EEI-TEL/0067/2012

CONTENTS

� Introduction

� Used System

� Cardiorespiratory signals

� Test procedure

� Cyclostationarity

� Results

� Conclusions

INTRODUCTION – SOFTWARE DEFINED RADIO

Software Defined Radio is a highly digital configurable radio that allows for the majority of processing to be done digitally

The processing needs

to be flexible and able

to be reconfigured in

real time.

Software Defined Radio

INTRODUCTION – COGNITIVERADAR

New generation of radars need to be aware of their environment, allowing:

� Usage of multiple radars without interference

� Intelligent usage of spectrum in order to get the best SNR

INTRODUCTION – BIO-RADAR

Heartbeat and breathing

pattern detected thanks to

doppler effect from the

reflection on the emitted

sinusoid

Chest movement can be used

to determine bio signals

USED SYSTEM - USRP B200 SDR

Shared clock at the Tx and Rx allow for Homodyne detectionallowing to avoid any fluctuations in the mixer frequency

USED SYSTEM - USRP B200 SDR

� CW with a frequency of2,7GHz

� Emited powerof 5mW

� Subject under test seated at 1meter

CARDIORESPIRATORY SIGNALS

� Bio-radar adquired signals for a person at rest

Typical values

in humans:

Repiration

0.1–0.8Hz

Heart Rate

0.8–2Hz

0 2 4 6 8 10-0.05

0

0.05Heart signal

Seconds

Am

plit

ude

0 2 4 6 8 10-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3Respiration signal

Seconds

Am

plit

ude

TEST PROCEDURE

0 20 40 60 80 100 120-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5

Seconds

Am

plit

ude

Start 1st test

(three breaths)

Apnea

(68bpm)

Normal Breathing

(68bpm)

Exercise

outside range

Start 2nd test

(three breaths)

Apnea

(76bpm)

Normal Breathing

(76bpm)

CYCLOSTATIONARITY

� A signal shows cyclostationarity if:� Auto-correlation function is periodic with time

� The mean also shows periodicity

� What signals are cyclostationary:� Biosignals like breathing and heartbeat show periodic features

Periodic auto-correlation

↓

Cyclostationarity

CYCLOSTATIONARITY FEATURES

� Cyclic spectrum heartbeat and breathing cyclostationarityfeatures

BreathingHeartbeat

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2-4

-3

-2

-1

0

1

2

3

4

Frequency

Cyclic

Fre

quency

-3 -2 -1 0 1 2 3-6

-4

-2

0

2

4

6

Frequency

Cyclic

Fre

quency

METHODS COMPARATION

� Energy detection by using peak of FFT

� Cyclostationaryby using cyclic spectrum peak

-25 -20 -15 -10 -5 0 50

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1Detection probability of the heartbeat under noise

SNR (dB)

Pro

babili

ty o

f dete

ction

Energy Detection

Cyclostationary features

� Improvement of 6dB for probabilities of detection higher then 80%

CONCLUSIONS

� It’s feasible to use a software defined radio to build a bio-radar and acquire respiration and heartbeat

� With a emitting power <10mW we can acquire the heartbeat from 1meter away and breathing for more than 5 meters

� Cyclostationary analysis improve the performance of this type of radars comparatively to simple energy detection schemes

QUESTIONS?