This article has been accepted for publication in a future issue of this journal, but has not been

fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JSEN.2018.2885048, IEEE Sensors
Journal
1

Accurate and Robust CW-LFM Radar Sensor:

Transceiver Front-End Design and Implementation
Homa Arab, Steven Dufour, Member, IEEE, Emilia Moldovan, Cevdet Akyel, Member, IEEE,
Serioja O. Tatu, Senior Member, IEEE

Abstract—This paper outlines the design and implementation of Working in the mm-wave range makes it possible to reduce the
a continuous-wave linear-frequency-modulated (CW-LFM) sensor size of sensor components, it leads to accuracy improvements,
for short-range radar applications. Low-cost technologies have and it makes measurements easier and cheaper to perform. Only
been selected to implement the sensor. It was integrated on a
0.127 mm thin ceramic substrate using Miniature Hybrid Mi- a few ISM bands are available between 24 and 24.25 GHz,
crowave Integrated Circuit (MHMIC) technology. Two 16 dBi 60 and 60.5 GHz, 75.5 and 81.5 GHz, 122 and 123 GHz, and
gain 8  2 patch antenna arrays are implemented in the trans- 244 and 246 GHz. The radar sensor industry is particularly
mitter and receiver modules. The receiver is a six-port homo- interested by the 5.5 GHz of allowed bandwidth around 77 GHz
dyne quadrature down-converter. The integrated power detectors and by the short wavelength owing to the fact that it pro-
are fabricated using HSCH-9161 Schottky diodes. The complete
sensor, in the metallic fixture, measures 35  45 mm. The per- vides high accuracy and resolution. Highly accurate distance
formance of the sensor to detect a target range and velocity is measurements are reported in [12]–[16], where an accuracy of
validated using system simulations and measurement systems on ˙ 2 m over a measurement range of 5 mm for a frequency of
a test bench. System simulations use computer models based on 122 GHz is obtained. This accuracy is attained for a very short
S-parameter measurement results. Measurements show that an measurement range over a limited 1 GHz bandwidth. In this
accuracy of ˙ 16.23 m can be achieved with the sensor over a
range of 15 cm. To precisely measure the six-port radar accuracy, work, an accuracy of ˙ 16.23 m over an electrical measure-
due to our RF laboratory capabilities, this result is obtained by ment range of 150 mm is achieved using a six-port radar. To
changing the electrical length, instead of varying the physical the best of our knowledge, the proposed radar sensor front-end
distance. is the first fully integrated 77 GHz prototype using Miniature
Index Terms—CW-LFM radar, homodyne, millimeter-wave, in- Hybrid Microwave Integrated Circuit (MHMIC) technology. In
phase==quadrature demodulator, reflectometer, sensor, six-port. addition to the aformentioned benefits, a large bandwidth makes
it possible to design high resolution frequency modulated radar
sensors. Six-port reflectometers make it possible to perform
I. I NTRODUCTION accurate phase measurements only using simple scalar mea-
Millimeter-wave radars based on a frequency-modulated surements. The six-port circuit has been introduced by F. Engen
continuous-waveform (FMCW) have become, in recent years, in the 20th century [17]. It was initially used, 28 years ago,
an increasingly widespread technology due to technical devel- as a two-port vector network analyzer (VNA) [18]. Six-port-
opments and to new application areas [1]-[5]. FMCW radars based devices are now being used in other fields for material
that use linear frequency modulation provide high-accuracy characterization, and for the design of receivers and radars
information about a target range and its relative velocity. They [19]–[25]. The proposed six-port reflectometer is composed of
also overcome the DC offset problem, often observed when a combination of four 90° hybrid couplers and of a 90° phase
using unmodulated CW radars [6]–[9]. CW-LFM sensors do shifter which is experimentally characterized and validated for
not suffer from transmission high peak-to-average power ratio millimeter-wave signal quadrature down-conversion.
(PAPR). This simplifies the design of antennas, power am- The sensor can measure phase, frequency and relative am-
plifiers and power detectors, when compared to pulse radars plitude variations of the reflected signal from the target, with
[10], [11]. Many industrial systems such as biomedical appara- respect to a reference signal derived from a 10 dB parallel line
tuses, construction equipments, fluid level gauging systems or coupler from the transmitted LFM-CW signal. For the proposed
robotics and automation systems rely on the determination of prototype, two 16 dBi gain 16-element patch antenna arrays
a distance or a position with high accuracy. Given the larger have been used in the transmitter and receiver modules. The
size and the low accuracy of low-frequency radar sensors, use of patch antenna arrays is justified by the fact that they
millimeter-wave radars are well-suited for these applications. are economical, compact, and easy to fabricate and set up.
The commercially available millimeter-wave GaAs Schottky
H. Arab received her Ph.D. degree from Institut national de la recherche diodes, Agilent HSCH-9161, are used to convert the output
scientifique (INRS). She is now a postdoctoral fellow at École Polytechnique
de Montréal, Canada. e-mail: [email protected].
power of the six-port to baseband voltages. For this project, the
S. Dufour is professor at École Polytechnique de Montréal, Canada. entire millimeter-wave RF front-end and T/R-modules of the
E. Moldovan is a research assistant at Institut national de la recherche LFM-CW radar are designed, fabricated and measured over a
scientifique (INRS), Montréal, Canada. wide bandwidth. In what follows, the benefits and limitations
C. Akyel is professor at École Polytechnique de Montréal, Canada.
S. Tatu is professor at Institut national de la recherche scientifique (INRS), of the proposed LFM-CW sensor will be outlined using system
Montréal, Canada. simulations and an appropriate measurement set-up.

1558-1748 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JSEN.2018.2885048, IEEE Sensors
Journal
2

TX Antenna
Line Coupler
R

Target
Six-Port Reflectometer Power Detectors SC.
DSP
V2

RX Antenna
DC Loop
70.7 Ω

V4
o
90
70.7 Ω

70.7 Ω
70.7 Ω

V1
70.7 Ω

V3
◦
90 Hybrid Coupler 70.7 Ω

Fig. 1: Block diagram of the six-port-based radar sensor.

II. T HEORY AND BASIC E QUATIONS TABLE I: Chirp parameters.

Parameter Value
The six-port-based FMCW sensor used for estimating the B 120 MHz
distance to a target using a multi-tone radar system will now T 650 s
be outlined. The block diagram of the radar is illustrated in f0 77 GHz
Fig. 1. The receiver is composed of a six-port reflectometer and
of diode-based power detectors to generate in-phase quadrature
(I/Q) differential signals. where m D B=T is the chirp rate, which can be seen as the rate
Various modulations (sawtooth, triangle, sinusoidal) are used of change of the frequency. Using the variation of the frequency
in the FMCW signals. In this case, we will consider a CW-LFM over time, we can find the instantaneous phase,
model of the FMCW signal, shown in Fig. 2. The transmitted Z t
f .t/ dt C '0 D 2 f0 t C m t 2 C '0 ; (2)


frequency increases linearly (in the bandwidth of B) as a func- 'TX .t/ D 2 2
0
tion of time during the sweep time T . For short range radar
applications, maintaining a wide bandwidth to reach a good where '0 is the initial phase of the signal. The transmitted signal
range resolution is critical. However, as the bandwidth gets can then be written as
wider, the noise level of the receiver will get higher. The values xTX .t/ D a cos 'TX .t/ :


(3)
of the chirp parameters are given in Table I. These parameters
are chosen based on our laboratory instruments resolution and To describe the transmitted signal in the nth sweep, we take
their finite accuracy for short range measurements. t D nT C ts , where ts is the initial time of the nth sweep. The
phase of the transmitted signal is then given by
'TX .t/ D 2 f0 .nT C ts / C m .nT C ts /2 C '0 : (4)


77.12
2
RF Signal Magnitude (dBm)

77.10 -30 The returned signal from the target will have the same form,
Frequency (GHz)

77.08 -50
but with a delay , which can be expressed as
 D 2c R C .nT C ts / v ;


(5)
77.06 B
... -70 120 MHz
where R is the range of the target, v is the target velocity, and
77.04 -90
c is the speed of the light. The returned signal can then be
77.02 -110 expressed as
T


77.00 xRX .t/ D ˛.t/ a cos 'RX .t/ (6)
0 0.6 1.2 77 77.15
where
Time (s) Frequency (GHz)
/ C m /2 C '0 : (7)


'RX .t/ D 2 f0 .nT C ts 2
.nT C ts
Fig. 2: Transmitted signal chirp modulation.
According to the theory of six-port reflectometers, the re-
turned signal is mixed with the transmitted signal. As shown in
The frequency is a linear function of time, with an initial
Fig. 3, the proposed six-port interferometer is a linear passive
value of 77 GHz, as illustrated in Fig. 2. It can therefore be
mm-wave circuit which is the combination of four 90° hybrid
expressed as
couplers (HC) and of a 90° phase shifter implemented using
f .t/ D f0 C mt; (1) microstrip transmission lines. The transmitted signal is injected

1558-1748 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JSEN.2018.2885048, IEEE Sensors
Journal
3

at port 5 as a local signal, and at port 6 as a received signal. The a short range object (less than 1 m) which is moving with
normalized waves at the four output ports of the six-port circuit a velocity of less than 10 m/s, equation (13) can then be
are approximated by
bi D a5 S5i C a6 S6i for i D 1; 2; 3; 4: (8) 
4 f0 R 
2f0 vn
I.t/  k a2 ˛.t/ cos 2 2mR c
t s C c
T C c ; (14)
1
2. j a5 a6 / where the first term is the beat frequency related to the range
4 C to the target, the second term represents the Doppler frequency,
Phase Shifter I
90ı HC 2 and the last term is the constant phase shift. This theoretical
blaı
90 1
2 .a5 C j a6 / study of the FMCW radar is the basis for the simulation and the
implementation of the platform of the developed circuit.
a5 (LO Input)
a6 (RF Input)
5 90ı HC 90ı HC 6 III. C IRCUIT D ESIGN AND R ESULTS
Various architectures for the six-port interferometer, which is
1
2. j a5 C j a6 / composed of a power divider, a coupler and a phase shifter, can
1 C Q be found in the literature [26]–[28]. Here, a six-port circuit with
90ı HC 3 four 90° hybrid couplers and a 90° phase shifter is experimen-
1 tally characterized and validated for RF signal demodulation.
2 .a5 C a6 /
This circuit has been fabricated on a 127 m thick ceramic
substrate, with a relative permittivity of "r D 9:8 using in-
Fig. 3: Layout of the six-port interferometer. house MHMIC technology. As shown in Fig. 4, the six-port RF
and local signal ports are well matched and the transmission
The four output ports of the six-port circuit are connected to between the RF input port and the four outputs is close to
power detectors that deliver a voltage that is linearly related to the theoretical value of 6 dB over a wide frequency band of
the power of the RF signal. We know that the output voltage of 12 GHz [27]–[29].
four identical power detectors is proportional to the square of
the magnitude of the RF input signal:
-5 -5
S S S S
S-Parameter Magnitude (dB)

Vi D kjbi j2 for i D 1; 2; 3; 4: (9) 61 62 63 64

-10 -6
The output voltages can be expressed as:
2

 -15 -7
V1;3 .t/ D ka4 1 C ˛ 2 .t/ ˙ 2˛ 2 .t/ cos '.t/ I

 (10) -20 -8
2
V2;4 .t/ D ka4 1 C ˛ 2 .t/ ˙ 2˛ 2 .t/ sin '.t/ ;
-25 -9
where ' D 'RX 'TX is the instantaneous phase difference S55
between the six-port inputs (LO and RF), and ˛.t/ is the in- -30 S -10
66
stantaneous amplitude ratio between RF and LO input signals.
Using the differential approach with equations (10), the output -35 -11
70 72 74 76 78 80 82
signals I and Q are given by
Frequency (GHz)
I.t/ D V3 .t/ V1 .t/

 (11) Fig. 4: Measurements of the S-parameter magnitude in dB.
m 2
D k a2 ˛.t/ cos 2 f0  C m ts 2
;
A successful six-port operation must accurately measure
and
high-frequency power ratios. HSCH-9161 Schottky diodes are
Q.t / D V4 .t/ V2 .t/ used in the designed power detectors to extract both the mag-

m 2

 (12) nitude and the phase information of the reflected signal from
D k a2 ˛.t/ sin 2 f0  C m ts 2
:
the target. Moreover, a 8  2 micro-strip patch antenna array is
Using equation (5) with equation (11) for , we get designed and fabricated for the transmitter and the receiver. In
 order to reduce the size of the whole prototype, and to ease the
I.t/ D k a2 ˛.t/ cos 4c f0 R C f0 vnT C f0 ts v C integration with other integrated passive devices, the proposed
mRts C mvnT ts C mvts2 antenna array has been designed in the same ceramic die. The

4 (13) simulated peak gains are 6.3 dBi for the element and 16.2 dBi
c2
mR2 C mv 2 n2 T 2 C for the array at 77 GHz.

 Fig. 5 shows the details of the fabricated six-port reflectome-
mv 2 ts2 C 2mvR nT C 2mvRts :
ter, of the power detectors and of the transmitter and receiver
Based on the parameters of Table I, the size of several terms antennas (all fabricated on a 16.83 mm  13.94 mm ceramic
of equation (13) are small and can therefore be neglected. For substrate) [30], [31]. The die is mounted and power detectors

1558-1748 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JSEN.2018.2885048, IEEE Sensors
Journal
4

are wire-bonded to the PCB using ribbon bonds to ensure good given in Fig. 7. It can be seen that these simulation results
matching over a broad bandwidth. for various Doppler frequencies also agree with the theoretical
values computed using the second term of equation (15), and
RX Antenna that all the parts are modeled accurately.
Since the front-end is optimized at a central frequency, the
output signal voltage amplitude slightly decreases if the fre-
Power Detector Six-Port
quency is shifted. Therefore the circuit also acts as a frequency
demodulator (FM to AM converter).
Measurements for various beat frequencies are illustrated in
TX Antenna Fig. 8. This additional property of the designed circuit can be
used to extract more accurate information about the target, at
higher beats or at Doppler frequencies.
The measurement set-up is composed of an HP-83550 series
millimeter-wave source, a millimeter-wave active multiplier
Fig. 5: The manufactured hardware prototype. from OML (X6), and a Tektronix DPO 7054 digital phosphor
oscilloscope. An Agilent N8241-A arbitary waveform genera-
To validate the proposed schematic, various parts of the tor (AWG) is used to generate chirp waveforms, based on the
circuit are fabricated separately, and measurement results are information of Table I, for creating LFM-CW signals using a
imported in the Advanced Design System (ADS) software Matlab interface. Matlab is one of the interfaces supported by
simulator of Keysight Technologies to build accurate computer the N8241-A. It facilitates the control and generation of output
models for the circuits used in system simulations. As a target, data. The AWG can create high-resolution waveforms for the
an aluminum plate with a cross-section of 0.01 m2 is placed at proposed millimeter-wave radar sensor.
a range of 0.3 m, 0.6 m, and 0.9 m. Fig. 6 shows the spectrum The millimeter source frequency is set to 12.83 GHz and
and waveforms of the baseband signals. Amplifiers with 20 dB the input power to 10 dBm. The multiplier generates a 77 GHz
gain are used at two IF output ports of the circuit for system signal with a power of 0 dBm at the input of the parallel
simulations. Simulation results for various ranges show that the line coupler. Preliminary experimental distance and frequency
baseband signal frequencies agree with the theoretical values measurements were performed in our laboratory. Fig. 9 shows
obtained from the first term of equation (15). the experimental set-up. Range measurement are performed in
Simulation results for various velocities of the target are two steps. A target (an aluminum plate with a cross section of

0 0
m1=369.24 (Hz) m1m2 m3 R=0.3 m m1 v=5 m/s
m1=2566.67 (Hz) m2 m3
-20 m2=738.48 (Hz) R=0.6 m -20 m2=5133.33 (Hz) v=10 m/s
m3=1107.72 (Hz) R=0.9 m m3=7700.00 (Hz) v=15 m/s
I Spectrum (dBm)
I Spectrum (dBm)

-40 -40

-60 -60

-80 -80

-100 -100
-10 -8 -6 -4 -2 0 2 4 6 8 10 -50 -40 -30 -20 -10 0 10 20 30 40 50

Frequency (KHz) Frequency (KHz)

0.3 0.3
R=0.3 m
0.2 R=0.6 m 0.2
I Magnitude (mV)

R=0.9 m
I Magnitude (mV)

0.1 0.1

0.0 0.0

-0.1 -0.1

-0.2 -0.2
v=5 m/s v=10 m/s v=15 m/s
-0.3 -0.3
0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.6 0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.6
Time (ms) Time (ms)

Fig. 6: Simulation results of basedband signal for various R. Fig. 7: Simulation results of basedband signal for various v.

1558-1748 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JSEN.2018.2885048, IEEE Sensors
Journal
5

0.25 given by [20].

4 R.f2 f1 /
fb D 20 kH z '2 '1 D: (15)
c
The measurement results for various frequencies are given in
-0.10 Fig. 10. As expected, all the points move around a circle with
0.2 a specific phase shift of 10:8°, which agrees with the expected
I Magnitude (mV)

target range given by equation (15). The center of the circle

fb D 30 kH z represents the DC-offset of the circuit, which depends on the
characteristics of the six-port circuit, on the differences between
the four power detectors at the output ports, and on the settings
of the oscilloscope if DC coupling is used.
-0.03
0.25 I (mV)

fb D 50 kH z DC Offset
2

-0.05
1
0 0.5 1.0 1.5 2.0 2.5
(0.08,0.12) Q (mV)
Time (ms) 0
-3 -2 -1 1 2 3

-1
Fig. 8: Simulation results for the I signal, for various beat
frequencies.
-2

-3
+ deg 16.23 micron

11.0
AWG N8241-A
10.9
Phase (deg)

PC 10.8

10.7
10.6

10.5
HP-83550 Multiplier OML (X6)
76.5 76.7 76.9 77.1 77.3 77.5

Frequency (GHz)
Target

Fig. 10: Measurement results for a target at R D 15 cm, in

Tektronix Oscilloscope steps of 30 MHz for frequencies between 76.5 and 77.5 GHz.

The phase difference for various ranges and frequencies are

V4 then measured. In Fig. 11, the frequency varies in steps of
V2 30 MHz for a bandwidth of 120 MHz. The phase difference
V3 77 GHz Sensor between two points give the target range. To get a better
V1 accuracy, larger frequency steps must be chosen for smaller
ranges, and the maximum phase difference must not exceed
360°. The DC-offset is removed to make the diagram easier to
Fig. 9: Short distance measurement set-up. read. This does not have any effect on target range and velocity
measurement results.
The area of the ellipse (a perfect circle in theory) is directly
proportional to the received power. The ratio of the received
0.09 m2 ) is first placed 15 cm from the fabricated sensor, and power to the transmitted power can be calculated using the radar
the phase shifts are measured for various frequencies in steps of equation,
f D 30 MHz from 76.5 GHz to 77.5 GHz. In fact, instead of   2  4     2
physically changing the length of the range, the electrical length Pr D Pt Gt Gr ; (16)
and the number of wavelengths was changed accordingly. It 4 R 2 4 R
would be difficult to make a calibrated moving mechanical where  represents the radar cross section (RCS), Gr is the
target as accurate. Both approaches to experimentally validate receiver gain, Gt is the transmitter gain, Pt is the transmitted
the model are equivalent. Expected phase shift for each step is power and Pr is the received power. This equation is valid for

1558-1748 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JSEN.2018.2885048, IEEE Sensors
Journal
6

I (mV)
R=10 cm
For the velocity measurements, the same plate is moved in
front of the sensor with three different velocities. The results
3
R=15 cm shown in Fig. 12 are extracted from the 50 ms screen data of the
2 R=20 cm oscilloscope display. The screenshots are not taken at the same
exact time and there are distortions in the sinusoidal output
signals. These distortions are due to the modulation produced
1 at the output of the six-port circuit. This is also observed in the
Q (mV) simulation results. The measurement results give the expected
-3 -2 -1 0 1 2 3
Doppler frequency values for various target velocities.
-1
References related to millimeter-wave radar sensors for one-
dimensional distance measurements in a single-target scenario
-2 are compared with the proposed radar sensor prototype in Ta-
ble III in terms of center frequency, bandwidth, range resolution
-3 and applications. It must be noted that each cited reference
describes a sensor designed to operate at a specific range and
Fig. 11: Magnitude of the I signal for various ranges, while accuracy for a given application. The transmitted power and
the frequency sweeps over 120 MHz. waveform, bandwidth, and accuracy are then selected accord-
ingly. As shown in table III, millimeter wave FMCW radar is
the mustly widely used in the recent short range radar sensors.
a point target with a RCS that is significantly smaller than This study describes the first six-port interferometer designed
the range [32]. For short range sensors and relatively large FMCW radar at 77 GHz on a 127 m ceramic substrate, and it
target sizes, the radar equation will not give accurate results is the first hardware design and implementation of a 77 GHz
in comparison with experimental measurement results. Based low-cost radar sensor made to obtain raw data without needing
on equation (16), the power has to increase by a factor of 16 six-port calibration. Since low millimeter-wave power is used
when the radius is doubled. However, the measurement results in the experiments, the results are obtained for short ranges
given in table II show that the received power is proportional (operating at 0 dBm for transmission, and less than 20 dBm
to 1=R2:47 for near field targets. Therefore, the measurements for the six-port reference LO signal). Laboratory equipment
allow us to conclude that for near field and large RCS, the use limitations led us to test the radar system for a bandwidth of
of the Friis equation, where the received power is proportional 120 MHz, which resulted in an accuracy of less than 20 m for
to 1=R2 , is a better approximation than the one provided by the the whole prototype (measurement based).
radar equation. An empirical coefficient was obtained in our The six-port-based receiver has a 12 GHz bandwidth with an
experiments. accuracy of 10 m, which is related to the circuit phase errors.
The WR12 to microstrip transmission line, and the diode-based
TABLE II: Measurement results for various ranges. power detector, also have a bandwidth around 12 GHz. The
Range (cm) Phase shift (deg) Radius (mV) Ellipse area (cm2 ) microstrip patch antenna array reduces the overall radar sensor
10 7.22° 3.09 9.55
bandwidth to around 6 GHz, covering the available ISM band
15 10.79° 2.17 4.71
between 75.5 GHz and 81.5 GHz.
20 14.45° 1.31 1.72

TABLE III: Comparison of millimeter-wave radar sensors.

Method Frequency Distance Bandwidth Accuracy Year Application Ref.
FMCW radar 24 GHz 17 m 3 GHz 800 m 2003 Industrial applications [33]
(mixer based) 24 GHz 50 mm 1 GHz 5 m 2015 Automotive radar applications [34]
61 GHz 100 mm 0.5 GHz 50 m 2014 High-accuracy applications [16]
80 GHz 2 mm 10 GHz ˙2 m 2015 Micromachining tools [15]
122 GHz 5/150 mm 1 GHz 2 / 250 m 2015 Hydraulic measurement systems [35]
FMCW radar 35 GHz 750 mm 5 GHz 120 m 1999 Liquid level measurements [36]
(six-port based) 24 GHz 12.5 mm 0.5 m 2012 Sensing applications [37]
77 GHz 11/37 mm 2.5 GHz ˙6ı 2012 Direction of arrival detectors [38]
24 GHz 1m 0.25 GHz 5 m 2013 Vital signs monitoring [39]
24 GHz 150 mm 1 GHz 150 m 2014 Piston position in a cylinder [12]
24 GHz 1.7 m 250 MHz 500 m 2014 Vibration monitoring [40]
Proposed design 77 GHz 150 mm 6 GHz 16.23 m 2018 Biomedical and industrial applications

1558-1748 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JSEN.2018.2885048, IEEE Sensors
Journal
7

10 -3(V) 10 -3(V) [4] G. Wang, C. Gu, T. Inoue, C. Li, ”A hybrid FMCW-interferometry radar
2 2 for indoor precise positioning and versatile life activity monitoring,” IEEE
100 Hz 200 Hz Trans. Microw. Theory Techn., vol. 62, no. 11, pp. 2812-2822, 2013.
1 [5] M. Jahn, R. Feger, C. Wagner, Z. Tong, and A. Stelzer, ”A four-channel 94-
I 0 GHz SiGe-based digital beamforming FMCW radar,” IEEE Trans. Microw.
0 Theory Techn., vol. 60, no. 3, pp. 861-869, Mar. 2012.
[6] T. Jaeschke, C. Bredendiek, S. Kuppers, and N. Pohl, ”High-precision
Q -2
-1 DBand FMCW-radar sensor based on a wideband SiGe-transceiver
MMIC,” IEEE Trans. Microw. Theory Techn., vol. 62, no. 12, pp. 3582-
-4 -2
3597, Dec. 2014.
0 25 50 0 25 50 [7] T. Baek, K. Dong-Sik, L. Sang-Jin, et al ”A transceiver module for FMCW
radar sensor using 94-GHz dot-type Schottky diode mixer,” IEEE Sensors
-3 J., vol.11, no.2, pp. 370-376, 2011.
10 (V) 10 -3 (V)
2 2 [8] Y. Xiao, C. Li, J. Lin, ”A portable noncontact heartbeat and respiration
monitoring system using 5-GHz radar”, IEEE Sensors J. J., vol. 7, no. 7,
pp. 1042-1043, Jul. 2007.
I 0 0 [9] A. Singh, V. M. Lubecke, ”Respiratory monitoring and clutter rejection
using a CW Doppler radar with passive RF tags”, IEEE Sensors J., vol. 12,
no. 3, pp. 558-565, Mar. 2012.
Q -2 -2 [10] E. Yavari, and O. Boric-Lubecke, ”Low IF demodulation for physiologi-
cal pulse Doppler radar, Microwave Symposium (IMS), IEEE MTT-S Int.,
300 Hz 400 Hz 2014.
-4 -4
0 25 50 0 25 50 [11] G. S. Woods, D. L. Maskell, and M. V. Mahoney, ”A high accuracy mi-
crowave ranging system for industrial applications,” IEEE Trans. Instrum.
Time (ms) Time (ms)
Meas., vol. 42, Issue(4), Aug. 1993.
[12] S. Lindner, F. Barbon, S. Linz, S. Mann, R. Weigel, and A. Kölpin,
Fig. 12: Measurement results for various velocity. ”Distance measurements based on guided wave 24 GHz dual tone sixport
radar,” in Proc. Eur. Radar Conf. (EuRAD), pp. 57–60, Oct. 2014.
[13] S. Lindner, F. Barbon, S. Mann, G. Vinci, R. Weigel, A. Koelpin, ”Dual
tone approach for unambiguous six-port based interferometric distance
The sensor circuit was tested for ranges between 10 and measurements,” In Proceedings of the IEEE MTT-S International Mi-
100 cm. The minimum range limit is the range at which the crowave Symposium, Seattle, WA, USA, pp. 1–4, 2–7 June 2013.
electromagnetic field is considered to be a plane wave. The [14] T. Jäeschke, M. Vogt, C. Bäer, C. Bredendiek, and N. Pohl, ”Improve-
ments in distance measurement and SAR-imaging applications by using
maximum range limit is related to the level of received power ultra-high resolution mm-wave FMCW radar systems,” in IEEE MTT-S Int.
which is close to the noise level at higher ranges. The designed Microw. Symp. Dig., pp. 1-3, Montreal, QC, Canada, Jun. 2012.
sensor can be used for longer ranges, using higher power levels [15] S. Ayhan et al., ”Millimeter-wave radar distance measurements in micro
machining,” in Proc. IEEE Topical Conf. Wireless Sensors Sensor Netw.
at the input, higher gain antennas, or low-noise amplifiers. (WiSNet), pp. 65-68, Jan. 2015.
The fabrication of all the components of the proposed sensor [16] S. Scherr, S. Ayhan, H. Gulan, M. Pauli, and T. Zwick, ”61 GHz ISM
circuit in the same die reduces the size of the circuit and the band FMCW radar for applications requiring high accuracy,” in Proc.
Asia–Pacific Microw. Conf. (APMC), pp. 1118-1120, Nov. 2014.
number of connections, resulting in reduced losses due to con-
[17] G. F. Engen, ”An Introduction to the Description and Evaluation of
nections and in a physically robust circuit. With this robust and Microwave Systems Using Terminal Invariant Parameters,” Nat. Bur. Stan-
low-cost schematic and manufacturing process, these sensors dards, Washington DC, 1969.
can be used in various industrial and biomedical applications. [18] A. L. De Souza Rolim, M. T. Belfort, ”Six-port Complex Permittivity
Measurements,” Proc. Eur. Microw. Conf. , pp. 492-494, Sept. 2006.
[19] S. O. Tatu, and K. Wu. ”Communication and Sensing Applications of
IV. C ONCLUSION Six-Port Technology” Microw. Review, vol. 19, Issue(2), 2013.
[20] E. Moldovan, S.O. Tatu, T. Gaman, Ke. Wu, R.G.Bosisio, ”A New 94
A novel millimeter-wave integrated radar sensor is proposed. GHz Six Port Collision Avoidance Radar Sensor,” IEEE Trans. Microw.
It can be used in FMCW or in multi-tone configurations to Theory Techn., pp. 751-759, 2004.
accurately measure the range and the velocity of a target. Given [21] F. M. Ghannouchi, A. Mohammadi, ”The Six-Port Technique,” Artech
House; Norwood, MA, USA: 2009.
the very low millimeter-wave power used in the experiments,
[22] M. Caron, C. Akyel, F. M. Ghannouchi, ”A Versatile Easy to Do six-port
the results are performed at short range. The use of a low-noise Based High Power Reflectometer,” J. Microw. Pow. Electromag. Energy,
amplifier in the receiver front-end should considerably increase pp 231-239, 1995.
the operating range. Given the robust and low-cost schematic [23] T. Hentschel, ”The six-port as a Communications Receiver,” IEEE Trans.
Microw. Theory Techn., pp.1039-1047, 2005.
and manufacturing process, such sensors can be used in various [24] J. M. Ávila-Ruiz et al, ”Ultra-Wideband VCO Phase-Noise Reduction
industrial and biomedical applications. Loop Based on a Six-Port Phase-Shifterless Delay-Line Discriminator,”
IEEE Trans. Microw. Theory Techn., vol. 62, no. 12, pp. 3076-3084, Dec.
2014.
R EFERENCES [25] Chi-Hsueh Wang, Hong-Yeh Chang, Pei-Si Wu, Kun-You Lin, Tian-Wei
[1] R. Feger, C. Pfeffer, W. Scheiblhofer, C. M. Schmid, M. J. Lang and A. Huang, Huei Wang, and Chun Hsiung Chen, ”A 60GHz Low-Power Six-
Stelzer, ”A 77-GHz Cooperative Radar System Based on Multi-Channel Port Transceiver for Gigabit Software-Defined Transceiver Applications,”
FMCW Stations for Local Positioning Applications,” IEEE Trans. Microw. IEEE Solid-State Circuits Conf. (ISSCC), pp. 192-194, 11-15 Feb. 2007.
Theory Techn., vol. 61, no. 1, pp. 676-684, Jan. 2013. [26] S.O. Tatu, E. Moldovan, Ke Wu, R.G. Bosisio, T. Denidni, ”Ka-Band
[2] J. Hasch, E. Topak, R. Schnabel, T. Zwick, R. Weigel, and C. Waldschmidt, Analog Front-end for Software Defined Direct Conversion Receiver,” IEEE
” millimeter-wave technology for automotive radar sensors in the 77 GHz Trans. Microw. Theory Techn., vol.53, no. 9, pp. 2768-2776, Sept. 2005.
frequency band,” IEEE Trans. Microw. Theory Techn., vol.60, no. 3, pp. [27] H. Arab., C. Akyel, S.O. Tatu, ”Wide-band Millimetre Wave Down-
845-860, Mar. 2012. converter Based on Six-port Circuit for Radar and Sensing Applications,”
[3] M. Pichler, A. Stelzer, P. Gulden, and M. Vossiek, ”Influence of sys- URSI Conf., Montreal, 2017.
tematic frequency-sweep non-linearity on object distance estimation in [28] H. Arab., S.O. Tatu, C. Akyel. ”Performance analysis for two different
FMCW/FSCW radar systems,” 33rd Eur. Microw. Conf., vol. 3, pp. 1203- structures of 77 GHz six-port correlator,” ANTEM Conf., 17th International
1206, Oct. 2003. Symposium on, IEEE, Montreal, 2016.

1558-1748 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JSEN.2018.2885048, IEEE Sensors
Journal
8

[29] H. Arab., S.O. Tatu, C. Akyel, ”Design and Characterization of a 77 GHz Homa Arab was born in Iran. She received her master’s degree in telecom-
Six-Port Modulator for an Automobile Radar,” Veh. Technol. Conf. (VTC- munications engineering from École Polytechnique de Montréal and her Ph.D.
Fall), IEEE 84th, Montreal, 2016. degree in telecommunications from Institut national de la recherche scien-
[30] D. Hammou, T. Djerafi, M. Nedil, S.O. Tatu, ”Design Considerations for tifique (INRS) – Centre Énergie Matériaux Télécommunications, Montréal,
On-wafer Millimeter Wave Measurements on Thin Ceramic Substrate,” Canada. She worked for the Moshanir Company, in the Control and Protection
IEEE Transactions on Measurements, vol.65, no.2, pp. 441-447, Feb. 2016. Department, from 2005 to 2011. She is now a postdoctoral fellow at École
[31] D. Hammou, M. Nedil, N. Kandil, E. Moldovan, S. O. Tatu, ” V- Polytechnique de Montréal. Her current research interests are microwave and
band millimeter-wave micro-strip to rectangular waveguide transition,” millimeter-wave circuit design, transmitter and receivers, radars, and dielectric
Microwave and Optical Technology Letters, vol.55, no.7, pp.1696-1700, measurements.
2013.
[32] R.Deban, H. Boutayeb, et al. ”Deterministic approach for spatial diversity
analysis of radar systems using near-field radar cross section of a metallic
plate,” IEEE Trans. Antennas Propag. vol.58, no.3, pp. 908-916, 2010.
[33] T. Musch, ”A high precision 24-GHz FMCW radar based on a fractional-
Steven Dufour received his B.Sc. degree in mathematics and his M.Sc. degree
N ramp-PLL”, IEEE Trans. Instrum. Meas., vol. 52, no. 2, pp. 324-327,
in applied mathematics from Université de Montréal, and his Ph.D. degree
Apr. 2003.
in engineering mathematics from École Polytechnique de Montréal. After
[34] S. Scherr, S. Ayhan, B. Fischbach, A. Bhutani, M. Pauli, T. Zwick, ”An
two years as an assistant professor in the Department of Mathematics of the
efficient frequency and phase estimation algorithm with CRB performance
University of Wyoming, USA, he returned to the Department of Mathematics
for FMCW radar applications,” IEEE Trans. Instrum. Meas., vol. 64, no. 7,
and Industrial Engineering at École Polytechnique de Montréal, now as an
pp. 1868-1875, Jul. 2015.
associate professor. He was also a visiting professor in Brazil (UERJ) and Saudi
[35] S. Scherr et al., ”Miniaturized 122 GHz ISM band FMCW radar with
Arabia (KAUST). He is teaching applied mathematics to engineering students
micrometer accuracy,” Proc. Eur. Radar Conf. (EuRAD), pp. 277-280, Sep.
and his current research interests are the development of new finite-element
2015.
based numerical methodologies for modeling fluid mechanics problems found
[36] A. Stelzer, C. G. Diskus, K. Lubke, H. W. Thim, ”A microwave position
in free surface and turbulent flows, for modeling electromagnetism phenomena
sensor with submillimeter accuracy”, IEEE Trans. Microw. Theory Techn.,
found in supraconductors and for the optimization of antennas, and for the study
vol. 47, no. 12, pp. 2621-2624, Dec. 1999.
of magnetohydrodynamics.
[37] F. Barbon, G. Vinci, S. Lindner, R. Weigel, A. Kölpin, ”A six-port in-
terferometer based micrometer-accuracy displacement and vibration mea-
surement radar,” IEEE MTT-S Int. Microw. Symp. Dig., pp. 1-3, Jun. 2012.
[38] G. Vinci, F. Barbon, et al., ”A wide-range 77 GHz direction of arrival
detector with integrated dual six-port receiver,” Microwave Symposium
Digest (MTT), Montreal, QC, Canada, Aug. 2012. Emilia Moldovan received her B.Sc. degree in electrical engineering from the
[39] G. Vinci, S. Lindner, et al., ”Six-port radar sensor for remote respiration Polytechnic University of Cluj-Napoca, Romania, in 1980, and her M.Sc.A.
rate and heartbeat vital-sign monitoring,” IEEE Trans. Microw. Theory and Ph.D. degrees in electrical engineering from École Polytechnique de
Techn., vol. 61, no. 5, pp. 2093-2100, May 2013. Montréal, Canada, in 2001 and 2006 respectively. From 1982 to 1997, she
[40] G. Vinci, S. Linz, et al., ”A six-port radar system for precise distance was a telecommunication engineer in the Quality of Service Department, at
measurements and vibration monitoring in industrial environments,” Sen- the National Company of Telecommunications, Rom-Telecom, in Bistrita-
sors and Measuring Systems 2014; 17. ITG/GMA Symposium, Nuremberg, Nasaud, Romania. Since 2006, Dr. Moldovan is now a research assistant at
Germany, Jun 2014. Institut national de la recherche scientifique (INRS) – Centre Énergie Matériaux
Télécommunications, Montréal, Canada. Her research interests include passive
microwave and millimeter-wave circuit design, wireless communications, radar
sensors, and imaging systems.

Cevdet Akyel was born in Samsun, Turkey. He received his Sup.Ing. degree
from the Technical University of Istanbul in 1971, where he also worked
as a research assistant and developed the first Turkish holograms in the
microwave laboratory, and wrote a book on their scientific applications. He
moved to Canada in 1972 to get his M.Sc.A. and D.Sc.A. degrees from École
Polytechnique de Montréal in 1975 and 1980 respectively. He had engineering
positions in 1974 and 1976 at Northern Telecom of Canada as a system engineer
in radio telecommunications. Since 1986, he has been a professor of electrical
engineering at École Polytechnique de Montréal where he teaches electromag-
netics theory and automated microwave instrumentation. In 1991, he joined
the Poly-Grames group involved in space electronics and advanced microwave
technologies. His main research interests are the permittivity and permeability
measurements with microwave active systems, the characterization of materi-
als (conductive polymers, superconducting ceramics, ferromagnetic materials,
etc.), and high-power microwave measurements and industrial applications.

Serioja Ovidiu Tatu (M’05, SM’13) received his B.Sc. degree in radio engi-
neering from the Polytechnic University of Bucharest, Romania, in 1989, and
his M.Sc.A. and Ph.D. degrees in electrical engineering from École Polytech-
nique de Montréal, Canada, in 2001 and 2004 respectively. He worked for the
National Company of Telecommunications, Rom-Telecom, in Bistrita-Nasaud,
Romania, as an RF engineer and Head of the Telecommunications Laboratory
from 1989 to 1993, and as a Technical Manager from 1993 to 1997. Dr Tatu is
presently full professor at Institut national de la recherche scientifique (INRS) –
Centre Énergie Matériaux Télécommunications, Montréal, Canada. His current
research interests are microwave and millimeter-wave circuit design, hardware
and software radio receivers, radars, and imaging systems.

1558-1748 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.