Vagle et al.

The Journal of Global Positioning Systems (2016) 14:4

DOI 10.1186/s41445-016-0004-6
The Journal of Global
Positioning Systems

ORIGINAL ARTICLE Open Access

Performance analysis of GNSS multipath

mitigation using antenna arrays
Niranjana Vagle* , Ali Broumandan, Ali Jafarnia-Jahromi and Gérard Lachapelle

Abstract
Multipath affects the shape of the correlation function and results in biased pseudorange measurements and erroneous
navigation solutions. Antenna array processing, which uses signal spatial characteristics, is an effective method to mitigate
various types of interference signals. However, the performance of most of the distortionless beamforming techniques
degrades in multipath conditions due to the correlation between the desired Line of Sight (LOS) signal and multipath
signals. This paper characterizes the performance of different beamforming techniques to mitigate multipath signals
through the processing and analysis of simulated and actual data. The main novelty is the investigation of multipath
mitigation performance of practically realizable antenna array-based GNSS receivers when the beamforming process is
completely integrated into the tracking module after de-spreading. Beamforming techniques such as Delay And Sum
(DAS) beamforming, Minimum Power Distortionless Response (MPDR) with and without spatial smoothing are
considered. A novel multi-antenna simulator test-bed is developed to generate multipath signals for a multi-antenna
platform. A software multi-antenna GPS receiver incorporating different beamforming techniques is then developed to
generate pseudorange measurements and position solutions. Carrier-to-Noise ratio (C/N0), pseudorange errors and
position solutions before and after beamforming are compared to show the effectiveness of different beamforming
techniques to mitigate multipath. Results with simulated and actual GPS signals show improved performance using the
MPDR beamformer with spatial smoothing. The utilization of spatial processing results in a pseudorange error reduction
of up to 60 % and a position error reduction of up to 30 %.
Keywords: GPS, Multipath, Beamforming, MPDR, Software simulator

Introduction Narrow Correlator™ (Dierendonck et al. 1992), Multipath

Although modern GNSS receivers provide high accuracy Estimating Delay Locked Loop (MEDLL) (Van Nee et al.
positioning and navigation solutions in open sky condi- 1994) and Edge Correlator (Garlin et al. 1996) to name a
tions, multipath remains a major error source in many few. However, multipath due to nearby reflectors is still
environments. Multipath results in a distorted correl- a major problem for correlator-based techniques.
ation function that is used to estimate delays and pseu- Antenna array processing, a signal processing scheme
doranges. This results in erroneous navigation solutions. that exploits the signal spatial features, is proven to be
Multipath also leads to incorrect ambiguity resolution effective in mitigating different types of interference.
affecting carrier phase positioning. If the multipath pseu- Even though antenna array processing is well studied for
dorange error becomes large, the initial position solution wireless communication systems, the application of
is biased and the carrier phase ambiguity search space these techniques to GNSS differ from those systems. For
can be enlarged, resulting in longer ambiguity resolution instance, in most wireless communication systems, in-
time (Joosten et al. 2002). Long-delay code multipath creasing the signal to noise ratio to reduce bit error rate
caused by distant reflectors can be mitigated using cur- is the main focus; for GNSS the focus is to improve
rently available advanced correlator techniques such as time-delay estimation to improve estimated position
accuracy. The effectiveness of different beamforming tech-
* Correspondence: [email protected]
niques for GNSS applications was studied in (Fern’andez-
Plan Group, Department of Geomatics Engineering, Schulich School of Prades et al. 2016; Gupta et al. 2016; Broumandan et al.
Engineering, University of Calgary, 2500 University Drive NW, Calgary, AB 2016; Cuntz et al. 2016; Amin et al. 2016; Daneshmand et al.
T2N1N4, Canada

© 2016 The Author(s). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made.
Vagle et al. The Journal of Global Positioning Systems (2016) 14:4 Page 2 of 15

2014; Arribas et al. 2014; Egea et al. 2014; Kalyanaraman and processing techniques and the decorrelation effect due
Braasch 2007; Kalyanaraman and Braasch 2010). Most of the to spatial smoothing are discussed in Section III. In
distortionless beamforming techniques are developed with Section IV, GPS multi-antenna signal simulation meth-
the assumption that there is no correlation between desired odology using a ray tracing method and beamforming
and interference signals. However, performance of these implementation is discussed. The results of multipath
beamforming techniques degrades in multipath interference mitigation using simulated signals are presented in Sec-
because there is a high degree of correlation between desired tion V and actual GPS signal processing results are pro-
and multipath signals (Van Trees 2002). The effectiveness of vided in Section VI. Finally, Section VII summarizes the
antenna arrays to mitigate multipath interference has been findings.
studied through different robust beamforming techniques
in GNSS applications (Brown 2000; Fu et al. 2003; Seco- Notation
Granados et al. 2005; Sahmoudi and Amin 2007; Throughout this paper, the following notations are
Konovaltsev et al. 2007; Vicario et al. 2010; Fern’andez- adopted: small bold letters represent vectors and capital
Prades et al. 2011; Daneshmand et al. 2013a; Manosas- bold letters represent matrices. Superscripts H and T
Caballu et al. 2013; Rougerie et al. 2011; Rougerie et al. represent complex conjugate transpose and transpose,
2012; Lee and Hsiao 2008). Sahmoudi and Amin (2007) respectively. A particular element in a rectangular array
used adaptive beamforming and high resolution direction is represented within parentheses as in (a, b), the subar-
finding methods to improve robustness against multipath ray is represented within square brackets as in [a, b] and
and electronic interference. Vicario et al. (2010) analyzed the subarray size is represented within braces as in {a,
robust beamforming techniques for Galileo ground sta- b}. The symbol a represents a quantity in the x-direction
tions and shown a reduction of tracking errors by 47 %. and b a quantity in the y-direction. The direction of the
Fernández-Prades et al. (2011) studied the inherent signal is represented as (Elevation, Azimuth).
capability of different eigen beamforming techniques to
mitigate multipath through simulations. Some of these Signal and system model
techniques assume either a linear array or a large planar Consider the case of a GNSS receiver equipped with an
array which is not however feasible for practical applica- M × N element uniform rectangular array. The elements
tions. Efficient maximum likelihood techniques to miti- are lying in the x-y plane and are spaced by dm in the x-
gate multipath are not practical for many applications due direction and dn in the y-direction as shown in Fig. 1.
to their high computational burden. Even though the re- The signals impinging on the antenna array are the de-
sults from the previous research have shown that effective sired signals, multipath and noise. For simplicity, signals
multipath mitigation is possible, the performance of an- from only one satellite are considered below. After
tenna array based GNSS receivers in terms of time-delay down-conversion and sampling, the digitized signal re-
estimation and position accuracy has not been analyzed ceived at the (m, n)th antenna element can be expressed
extensively. Such performance is therefore assessed herein as (Van Trees 2002)
in terms of measurement and position accuracy through
X
K
different beamforming techniques. xm;n ðnt Þ ¼ sk ðnt Þej λ ½ðm−1Þdm sinðθk Þ sinðϕ k Þþðn−1Þdn sinðθk Þ cosðϕ k Þ
2π

The focus is on short-range multipath signals with k¼1

specular reflections. As GNSS signals are below the noise þvmn ðnt Þ
level before the correlation process, spatial processing to ð1Þ
mitigate multipath signals is mostly performed after the
de-spreading process (i.e., correlation and Doppler re-
moval) (Arribas et al. 2012; Chen et al. 2012). The inher-
ent capability of DAS and MPDR beamformers to
mitigate multipath are studied first without any prepro-
cessing to decorrelate the LOS and multipath signals. A
preprocessing technique called spatial smoothing is used
to decorrelate the signals. This process consists of two
stages. In the first stage, spatial smoothing is used later to
decorrelate LOS and multipath signals while in the second
stage, spatially smoothed signals are combined using the
MPDR beamformer. Measurement and position results
from simulated and actual GPS signals are provided.
The system model and the main assumptions are out-
Fig. 1 Uniform rectangular array configuration
lined in Section II. Effects of multipath on antenna array
Vagle et al. The Journal of Global Positioning Systems (2016) 14:4 Page 3 of 15

where sk ðnt Þ is the kth signal component observed at the h iT

ðN−1Þ
ak ¼ bTk ; γ k bTk ;::::::; γ k bTk
antenna element, k = 1 refers to the desired signal and k h i
ðM−1Þ T
= 2:K refers to multipath signals, λ is the signal wave- bk ¼ 1; βk ;::::::βk
length, ðθk ; ϕ k Þ are the elevation and azimuth angles of 2π
the kth component, vm;n ðnt Þ is the additive spatially white j dn sinðθk Þ cosðϕ k Þ
γk ¼ e λ
noise of the (m, n)th antenna element, and nt represents 2π
the discrete time index. j dm sinðθk Þsinðϕ k Þ
In this research, multipath mitigation is performed βk ¼ e λ
after the de-spreading process. Hence, the signal model ð9Þ
after the correlation process is considered here. Let the Digital beamforming solutions
correlator output signal for the (m, n)th antenna element This section describes the two different beamforming
be expressed by solutions considered in this research, namely DAS and
MPDR with and without spatial smoothing. The effect of
X
K
correlation between LOS and multipath signals on
rk ðnc Þej λ ½ðm−1Þdm sinðθk Þ sinðϕ k Þþðn−1Þdn sinðθk Þ cosðϕ k Þ
2π
ym;n ðnc Þ ¼
k¼1
beamformers is discussed and different numerical simu-
þηm;n ðnc Þ lations are performed to evaluate the performance of
these beamforming techniques to mitigate multipath
ð2Þ signals for GNSS applications. The main difference be-
tween GNSS and other systems is that the measurement
where nc represents the time index after correlation,
quality is of utmost importance beside signal strength
ηm;n ðnc Þ is the white noise component and rk ðnc Þ shows
improvement. Any type of filtering that distorts meas-
the correlator output of the kth signal component ob-
urement quality affects GNSS receiver performance.
served at the (1, 1) antenna element and is given by
Hence, special care is required for beamforming design
and implementation.
rk ðnc Þ ¼ αk ej2πΔfk nc Tc þjΔφk ð3Þ

where αk is the attenuation factor, Δfk represents the fre- DAS beamformer
quency offset and Δφk is the phase shift; Tc is the coher- The DAS beamformer relies only on the spatial informa-
ent integration time. tion of the LOS signal (Van Trees 2002). This beamfor-
The correlator output from all the antenna elements mer does not guarantee a distortionless response as it
can be represented in matrix form as just points the main beam in the direction of the LOS
signal and does not consider any other constraints to
y ¼ Ar þ η ð4Þ preserve the desired correlation peak shape. From Eqs.
(8) and (9), the steering vector of the LOS signal is given
where y is the MN × 1 correlator output vector, A is the by a1 . The optimum weights for the DAS beamformer
steering matrix, is MN × 1 noise vector, r is K × 1 corre- can be obtained as
lator output vector; these vectors can be written as 1
wCONV ¼ a1 ð10Þ
 T MN
y ¼ y1;1 ðnc Þ ; y2;1 ðnc Þ; …yM;1 ðnc Þ; y1;2 ðnc Þ; … yM;N ðnc Þ
where MN is the total number of antenna elements in
ð5Þ
the array.
h iT
η¼ η1;1 ðnc Þ; η2;1 ðnc Þ ; …ηM;1 ðnc Þ; η1;2 ðnc Þ; … ηM;N ðnc Þ MPDR beamformer
ð6Þ The MPDR beamformer is a distortionless beamformer
that minimizes total output power by constraining unity
r ¼ ½r1 ðnc Þ ; r2 ðnc Þ ; …rK ðnc Þ T ð7Þ gain in the direction of the desired signal (Van Trees
2002). This beamformer relies on the covariance matrix
The steering matrix A is of dimension MN × K is and of the received signal, which is normally computed by
given by temporal averaging of the spatial samples. The covari-
ance matrix of the received signal can be obtained as
A ¼ ½a1 ; a2 ; ::::::aK  ð8Þ
1 X KT
Ryy ¼ yyH ð11Þ
where ak is the MN × 1 steering vector of the k sig- th
KT k¼1
nal component coming from direction ðθk ; ϕ k Þ and is
given by where KT is the number of temporal snapshots.
Vagle et al. The Journal of Global Positioning Systems (2016) 14:4 Page 4 of 15

The optimum weight vector for the MPDR beamfor- MPDR beamformer with spatial smoothing (MPDRSS)
mer is (Van Trees 2002) Consider an M × N array divided into overlapping subar-
rays of size {J,L}. Assume P subarrays in the x-direction
and Q in the y-direction. Let Rfpq be the covariance
R−1
yy a1
wMPDR ¼ ð12Þ matrix of the [p, q]th forward subarray. The forward
−1
aH
1 Ryy a1 spatially smoothed covariance matrix is the sample
means of all the forward subarray covariance matrices
and can be computed as

Effect of multipath signals on beamforming 1 XP X Q

Rf ¼ Rfpq ð13Þ
The correlation between LOS and multipath signals has an PQ p¼1 q¼1
adverse effect on the beamformer’s performance (Widrow
et al. 1982; Reddy et al. 1987; Daneshmand et al. 2013b). Similarly, if Rb is the backward spatially smoothed co-
As the covariance matrix is obtained by temporal aver- variance matrix, then the forward-backward spatially av-
aging, the temporal cross correlation between the desired eraged covariance matrix is given by
and the multipath signals is very high since their phase re-
lation stays fairly constant during the averaging time. Rf þ Rb
Rfb ¼ ð14Þ
Therefore, the system regards the sum of the desired and 2
multipath signals as one wave and computes weights to The optimum weight vector for the MPDR beamfor-
minimize the total output power. However, as desired and mer with spatial smoothing is (Van Trees 2002)
multipath signals are treated as one wave, the weights will
have a destructive effect on the desired signal and in the R−1fb a11
wMPDRSS ¼ ð15Þ
process of mitigating multipath, the desired signal will also a11 R−1
H
fb a11
be cancelled (Widrow et al. 1982). In addition, the beam-
former fails to form deep nulls in the direction of multi- where a11 is the steering vector of the LOS signal for the
path (Chen et al. 2012). If the phase relation between the first subarray.
desired signal and multipath can be randomized, then the Beamformer’s performance depends on a number of
coherence between the signals will be reduced. This can be factors such as the number of antenna elements, array
achieved by receiving the signals from different spatial configuration and incoming signal directions of arrival
locations by the antenna array; this can be performed ei- to name a few. The size and number of antenna ele-
ther via moving the array (Daneshmand et al. 2013b) or ments are some of the limitations for practical applica-
through spatial smoothing techniques (Reddy et al. 1987). tions in terms of cost and system complexity. Hence
In the case of a static GNSS receiver, spatial smoothing can investigation of the performance of an antenna array
be applied to decorrelate the signals. In this method, an- based GNSS receiver with a limited number of antenna
tenna elements are grouped into a smaller number of over- elements while still being able to perform spatial
lapping subarrays (Van Trees 2002; Reddy et al. 1987). The smoothing is important. In this research a Uniform
basic requirement for spatial smoothing is that the steering Rectangular Array (URA) with six antenna elements is
vector should have a Vandermonde structure as in the case considered (M = 3, N = 2). The subarray formation for
of linear and rectangular arrays (Van Trees 2002). The the spatial smoothing is shown in Fig. 2. Due to the lim-
Vandermonde structure refers to the progressive linear ited number of elements in the array, only two subarrays
phase shift of the signals across the array elements. The (P = 2, Q = 1) are constructed with size {J = 2, L = 2}. The
covariance matrices from all the subarrays are then decorrelation obtained by spatial smoothing and in turn,
averaged to form the spatially smoothed covariance the performance of the beamformer, is analyzed in the
matrix. The subarray concept emulates antenna array following sections.
motion where signals received by different subarrays
correspond to different spatial points. In this case, Numerical simulations
the phase relation between LOS and multipath is dif- This section presents numerical simulation results for
ferent for different subarrays and averaging the spatial the array structure shown in Fig. 2 with inter-element
covariance matrix over several subarrays reduces the spacing of 9.5 cm. The performance of the beam-
correlation between the LOS and multipath signals. forming techniques in the presence of multipath sig-
Along with forward smoothing, complex conjugated nals is evaluated using the Signal-to-Multipath Ratio
backward smoothing can be performed to improve (SMR) (Egea et al. 2014) metric. SMR refers to the
the decorrelation as well as increase the antenna ratio between the LOS power and multipath power at
aperture (Reddy et al. 1987). the output of the beamformer and is expressed in dB.
Vagle et al. The Journal of Global Positioning Systems (2016) 14:4 Page 5 of 15


2

Rs ¼ σ r1 σ r1 σ r2 ρ
2
ð17Þ
σ r1 σ r2 ρ σ r2

where σ 2r1 is the variance of the source signal, σ 2r2 is the

variance of the multipath signal and ρ is the correlation
coefficient between the LOS and multipath, defined as
 
Ε r1 rH
ρ ¼ rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2
ð18Þ
 H qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 H
Ε r1 r1 Ε r2 r2

The power of both LOS and multipath is set to 10 (σ 2r1

¼ σ 2r2 ¼ 10 ) and the noise variance is assumed to be 1.
The elevation of the LOS signal varies from 0° to 90° for
Fig. 2 Subarray architecture for a 3 × 2 URA different magnitudes of the correlation coefficient be-
tween the signals and the SMR performance of both
MPDR and MPDRSS is shown in Fig. 3. The MPDR per-
The pre-beamformer SMR is assumed to be 0 dB. formance is the same for different LOS signal elevations
Here, it is assumed that multipath is coming from for a given correlation coefficient. For very low correl-
(15°, 175°) and the LOS signal azimuth is (50°). Beam- ation coefficients, which is the case when LOS and mul-
former performance for different correlation coeffi- tipath signals are uncorrelated to each other, the MPDR
cients of the LOS and multipath signals for different beamformer yields a SMR up to 40 dB. However, as cor-
LOS signal elevations is assessed. For the two signals relation increases, beamformer performance decreases
case, r1 (LOS) and r2 (multipath) the covariance and results in low SMR. As seen in Fig. 3, when the cor-
matrix can be represented as relation coefficient magnitude is above 0.6, the SMR is
nearly 0 dB. The performance of MPDRSS is better for
Ryy ¼ ARs AH þ σ 2η I ð16Þ higher elevation satellites when signals are correlated to
each other, as compared to MPDR. This is due to the
fact that the angular separation of the LOS from multi-
where Rs is the source covariance matrix and σ 2η is the path signals is higher and spatial smoothing is able to
noise variance. The source covariance can be defined as provide better decorrelation. As can be seen in Fig. 3,

Fig. 3 Output SMR performance with MPDR and MPDRSS with multipath coming from (15°, 175°) and LOS azimuth (50°) for different magnitudes
of correlation coefficient. [Indicates improvement in SMR for higher elevation satellites using MPDRSS as compared to MPDR as correlation
between signal increases]
Vagle et al. The Journal of Global Positioning Systems (2016) 14:4 Page 6 of 15

SMR up to 10 dB can be achieved using MPDRSS for the correlation between LOS and multipath is very
higher elevation satellites even when signals are highly low. However, MPDRSS provides better attenuation of
correlated. Since the decorrelation achieved by the the multipath signals. Even when signals are highly
spatial smoothing process is a function of the DOA of correlated, MPDRSS can attenuate multipath by up to
the incoming signals and the number of antenna ele- 10 dB. Based on the LOS signal directions and correl-
ments, the MPDRSS beamformer performance will be ation between LOS and multipath signals, the DAS
different for different signals impinging on the array beamformer performance could be similar to that of
from different directions. However, it was observed that MPDR and MPDRSS. In some cases, it could be better
for the rectangular array considered, MPDRSS beamfor- than MPDR as correlation can degrade the perform-
mer performance improves with an increase in the eleva- ance of the latter.
tion angle of the LOS signal, considering the multipath
signal is coming from a low elevation. Methodology
The beampatterns for the DAS, MPDR and MPDRSS This section describes the multi-antenna GPS signal
beamformers for different correlation coefficients for a simulator and receiver architecture used for the analysis
higher elevation satellite with multipath from low eleva- in multipath environments.
tion are shown in Fig. 4. Here it is assumed that LOS is
coming from (75°, 50°) and multipath from (15°, 175°). Multi-antenna GPS signal simulator
As the DAS beamformer does not rely on the statistics The multi-antenna GPS signal simulator can simulate
of the received signal, the performance will be same for GPS signals for a given user scenario and antenna array
any correlation between LOS and multipath signals. configuration. It has the option to simulate different
However, MPDR performance is improved only when multipath signals utilizing a ray-tracing approach. The

Fig. 4 Beampatterns with LOS (70°, 50°) and multipath (15°, 175°) a DAS beamformer b MPDR beamformer c MPDR beamformer with spatial
smoothing. [Better multipath attenuation occurs with the MPDRSS beamformer]
Vagle et al. The Journal of Global Positioning Systems (2016) 14:4 Page 7 of 15

main advantage of a software simulator as compared to vectors are computed for all satellites. The replica LOS
the use of actual data is the ability to control error signal is then multiplied by the LOS array steering vec-
sources such as antenna calibration uncertainties, at- tor to generate multi-antenna signals.
mosphere, multipath and clock errors. Therefore, the Based on the reflector and the satellite positions, the
performance of a beamformer can be analyzed for differ- point of reflection for the multipath signals is computed.
ent multipath signals parameters. The basic blocks of Once the reflection point is found, the extra distance trav-
the simulator are shown in Fig. 5. The input is the digi- elled by multipath signals is converted to the number of
tized IF samples collected using a data acquisition sys- chips, which is then added to the LOS prompt code to
tem either from a hardware similator or actual signals. generate multipath signals. Due to the additional path
These digitized samples are free of multipath. The soft- travelled by these signals and the reflection location, the
ware simulator is configured through two option files. Doppler observed by a multipath signal will be different
The main option file defines the parameters such as from that of the LOS signal. The multipath signal SMR
sampling frequency, channel numbers and satellite list. for different satellites is defined in the multipath option
The second option file is related to multipath signal pa- file. Using multipath Doppler information, replica code
rameters and defines the number of reflectors, reflector and the attenuation factor, multipath signals are generated
coordinates, user motion scenario and the antenna array for each visible satellite. Using the point of reflection and
configuration. known user position, The DOAs of multipath signals are
Using the digital samples, visible satellites are acquired computed and the corresponding steering vectors are gen-
and tracked. During the initial state of tracking, satellites erated using Equation (9) The multi-antenna multipath
are tracked in Phase Locked Loop (PLL) with higher signals thus generated for a particular satellite are then
bandwidth and loop order without assisting Delay added to the corresponding LOS multi-antenna replica
Locked Loop (DLL). Later, based on the Phase Lock In- signals. The combined LOS and multipath signals from all
dicator (PLI), the tracking state is switched to the PLL- visible satellites are added to generate the composite GPS
assisted DLL mode. The replica signals from this stage baseband signal. Later, independent noise is added to each
are used to generate multi-antenna signals. The replica antenna signal to have the desired C/N0 values for the
signal consists of code replica, carrier replica and the simulated signals.
navigation data bits. Using ephemeris information, satel- In order to evaluate the performance of the proposed
lite positions are computed. The satellite DOA is then multi-antenna software simulator, the IF sample files gen-
computed using these satellite positions and the known erated from the software simulator (reference antenna IF
user position with accuracy of a few decimetres or bet- file) and the Spirent hardware simulator were processed
ter. Based on the antenna array configuration defined in with the GSNRx software receiver (Petovello et al. 2008).
the option file and satellite signal DOAs, LOS steering The carrier Doppler values from the software receiver

Fig. 5 Multi-antenna GPS software simulator

Vagle et al. The Journal of Global Positioning Systems (2016) 14:4 Page 8 of 15

Fig. 6 Multi-antenna GPS receiver implementation

were inter-compared and similar performance was ob- combined correlator arms, namely early-prompt-late, are
served in terms of C/N0, signal tracking and navigation so- used by the tracking loops to generate the code and car-
lution. The validation process showed that the rier replica signals.
performance of the multi-antenna GPS software simulator A narrow correlator approach with 0.1 chip spacing be-
is comparable with that of a hardware simulator. tween early and late arms and a normalized non-coherent
early minus late envelope code discriminator are used. A first
Multi-antenna GPS receiver order DLL with bandwidth of 0.1 Hz is used in the PLL-
An open source single antenna MATLAB™ based GPS assisted DLL mode. The C/N0 is computed using narrowband
software receiver (Borre et al. 2007) was modified for power and wideband power as described in (Dierendonck
multi-antenna receiver functionalities. The acquisition, 1996). The least squares method is used to compute the pos-
tracking and navigation strategies of the original software ition solution with pseudorange measurements.
receiver were modified. The basic blocks of the multi-
antenna receiver are shown in Fig. 6. One of the antenna Results and discussions
elements in the array acts as the reference antenna. Satel- Simulated data
lite signals are acquired and tracked using the digital sam- This section describes the multi-antenna GPS signal simu-
ples of the reference antenna. The Doppler and code lation scenarios and the corresponding results for different
delays thus obtained are used to despread the signals from beamforming techniques. A GPS receiver equipped with a
other antennas so that relative phase values between the rectangular array as shown in Fig. 2 is considered for the
antenna elements are maintained. After Doppler and code simulations. A static user scenario was generated using a
removal from the digital samples corresponding to each Spirent hardware simulator; atmospheric, satellite clock
antenna, the prompt correlator values are used to com- and multipath errors were disabled. The GPS signal from
pute the optimum weights using the MPDR beamformer. the hardware simulator was sampled at 20 MHz using a
In order to capture the statistics of the incoming signals, National Instruments (NI) data acquisition system, which
prompt correlation values collected over one second are is input to the multi-antenna GPS signal simulator. Four
used to compute the covariance matrix of the MPDR rectangular shaped reflectors with dimensions of 30 m ×
beamformer. Thus its weights are updated every second. 50 m were considered. The reflectors were placed at a
The DAS beamformer does not use the statistics of the
prompt correlation values as it relies only on the satellite Table 1 Satellite DOAs used in simulations
DOA. Weights for the DAS beamformer are also updated PRN 6 10 16 18 21 24 25 29 31
every second to capture the LOS signal DOA variations. Azimuth (degrees) 281 34 280 149 119 57 320 91 211
The weights computed are used to combine 1 ms, early,
Elevation (degrees) 32 13 63 21 79 39 15 32 8
prompt and late correlator values of the six antennas. The
Vagle et al. The Journal of Global Positioning Systems (2016) 14:4 Page 9 of 15

Table 2 Receiver software execution configurations data for pseudorange error analysis. The pseudorange er-
Mode Configuration rors are computed by taking the differences between the
Mode 1 LOS scenario is assumed. Reference antenna tracks all of pseudoranges obtained in Mode 1, which is the reference
the observable signals. scenario, with those of Mode 2 to 5. The C/N0 values
Mode 2 LOS and multipath scenario is assumed. Reference antenna and pseudorange errors for various mitigation scenarios
tracks all of the observable signals. (Mode 2–5) in the case of PRN 6 are shown in Fig. 7. In
Mode 3 LOS and multipath scenario is assumed. Multi-antenna receiver this scenario the reflector-receiver distance was 30 m.
tracks all the observable satellites utilizing DAS beamformer. Periodic variations can be observed in the C/N0 values
Mode 4 LOS and multipath scenario is assumed. Multi-antenna receiver of Mode 2 due to the presence of multipath signals.
tracks all the observable satellites utilizing MPDR beamformer. Similar C/N0 fluctuations were also observed in other
Mode 5 LOS and multipath scenario is assumed. Multi-antenna receiver similar measurements (Ray et al. 1999). After beamform-
tracks all the observable satellites utilizing MPDR beamformer
with spatial smoothing process.
ing with the six antennas (Mode 2 to 5), the C/N0 varia-
tions are reduced and improvements occur. The C/N0
values improve by 8 dB in Mode 3 and 4 and 6.5 dB in
30 m distance from the user in all the four directions. The Mode 5 as compared to Mode 1. The reason that C/N0
reason for selecting reflectors in all the four directions is values in Mode 5 are less than those of Mode 3 and 4 is
to simulate multipath for most of the low elevation satel- because a lower number of antennas is used during
lites. Only specular multipath is considered with single re- beamforming due to spatial smoothing process. Simi-
flection. A multipath attenuation factor of 0.75 was larly, pseudorange errors after beamforming, which are
considered for each of the multipath signal. The DOAs of correlated to C/N0 variations, are significantly reduced,
different satellites used in the simulation are tabulated in indicating mitigation of the multipath signal using all
Table 1. Based on the ray tracing method, PRN 16 and three beamforming techniques.
PRN 21 do not observe any multipath. Comparisons of C/N0 and pseudorange RMS errors
The performance of the beamformer was evaluated by for all PRNs before and after beamforming are shown in
analyzing the improvement in C/N0 and multipath error Fig. 8. It is observed that the average C/N0 gain for all
reduction before and after beamforming. “Before beam- the satellites is the same for each beamformer. The gain
forming” refers to the tracking results obtained using obtained using the MPDR beamformer with spatial
baseband samples from the reference antenna. The smoothing is lower than that of the other two due to the
multi-antenna software receiver is executed in five dif- lower number of elements used in the beamforming
ferent configurations to generate C/N0 and pseudorange process. The pseudorange error reduction is different for
observations as described in Table 2. different PRNs. The MPDR beamformer with spatial
The received signal in Mode 1 is not affected by multi- smoothing provides better attenuation of multipath sig-
path and hence can be considered as a reference clean nals than the other two. For all the three beamformers,

55
Mode 1
50 Mode 2
C/N0 (dB-Hz)

45
Mode 3
Mode 4
40 Mode 5
35

30
0 20 40 60 80 100 120 140 160 180 200
40
Mode 2
Mode 3
PR error (m)

20
Mode 4
0 Mode 5

-20

-40
0 20 40 60 80 100 120 140 160 180 200
Time (s)
Fig. 7 C/N0 values and pseudorange errors for PRN 6
Vagle et al. The Journal of Global Positioning Systems (2016) 14:4 Page 10 of 15

50
Average C/N0 (dB-Hz)

45

40

Before beamforming
DealyAndSum
MPDR
MPDRSS
35
31 10 25 18 6 29 24 16 21
PRN
18

16

14

12
PR error (m)

Before beamforming
10
DealyAndSum
8
MPDR
MPDRSS
6

0
31 10 25 18 6 29 24 16 21
PRN
Fig. 8 C/N0 and pseudorange error performance comparison before
and after beamforming

it can be observed that, for very low elevation satellites Fig. 9 Live data collection a Setup showing antenna array and data
(<15°) such as PRN 10, 25 and 31, the pseudorange error collection system b Location of data collection and sky plot
reduction is minimal as compared to that for satellites
located at a higher elevation. This can be justified as the
signal decorrelation depends on the angle of arrival of 2016) arranged in a rectangular fashion with 11 cm spa-
the LOS and multipath signals. Since decorrelation has a cing between them. The array was mounted on a
direct impact on the performance of beamformer, the at- wooden platform on one end and a Novatel SPAN™ LCI
tenuation of the multipath signals by the beamformer inertial system was mounted on the other end to provide
also depends on the direction of arrival of the signals. platform attitude. Signals from the antenna array were
Considering Fig. 8b, DAS and MPDR beamformers can collected using a Fraunhofer multiple RF front-end,
reduce multipath errors by 2 to 8 m, whereas the which can collect digital samples from all the antennas
MPDRSS beamformer can reduce the errors up to 13 m. simultaneously. The location of the data collection and
The MPDRSS multipath reduction performance is much the corresponding sky plot are shown in Fig. 9. The glass
better than other techniques for all PRNs. building on the east side of the location acts as a specu-
lar reflector to generate multipath signals for the low
Field-test results elevation satellites visible in the west direction. Most sat-
GPS data was collected in moderate specular multipath ellite signals on the east side of the data collection loca-
conditions. The location was chosen such that both LOS tion were blocked by the building.
and multipath signals were observable with LOS being In order to perform array calibration, another data set
stronger than multipath signals. The setup, shown in was collected in open sky conditions with minimal mul-
Fig. 9a, consists of six NovAtel 501 antennas (Novatel tipath effect. The tracking architecture described in Fig. 6
Vagle et al. The Journal of Global Positioning Systems (2016) 14:4 Page 11 of 15

excluding the beamforming process was used to obtain The second analysis performed shows the improve-
the prompt correlation values to perform calibration. Car- ment in C/N0 values and pseudorange error reduction
rier and code were tracked by the reference antenna and after the beamforming process. A modified multi-
passed to the carrier and code tracking loops of the other antenna software receiver was used for this analysis. The
antennas. The Doppler and code replica signals obtained C/N0 values before and after beamforming were com-
after tracking the reference antenna signals are used to puted for different PRNs and the results are shown in
track other antenna signals to obtain relative signal ampli- Fig. 11a. “C/N0 before beamforming” refers to the C/N0
tude and phase values between different antennas. The computed from the reference antenna signal. Consider-
prompt correlator values of all the antennas were used to ing PRN 17, which is affected by multipath, the varia-
construct the steering vector, which is referred to as the tions are reduced after combining signals from all
measured steering vector. Based on the attitude of the antenna elements through beamforming and a 4 to 8 dB
array and DOA of the satellite, the true steering vector improvement is obtained. All three beamformers are
was computed. A least squares based calibration method able to reduce C/N0 variations.
(Backén et al. 2008) was used to compute the calibration To evaluate the pseudorange multipath error reduc-
matrix. As the number of visible satellites was larger than tion, a closely spaced base station with known position
the number of antenna elements, very low elevation satel- was used. A Novatel Propak receiver was used to collect
lites were excluded from the calibration process to avoid data at the base station. By using the ephemeris informa-
calibration errors due to multipath. tion and the user position, the true range could be com-
The initial analysis shows that some of the satellites are puted for each PRN. The pseudorange is the sum of true
disturbed by multipath signals. An independent variation range and other errors such as ionospheric, tropospheric
of C/N0 values from different antenna elements confirms and satellite clock errors, and multipath and measure-
the existence of multipath (Brown 2000). The C/N0 values ment noise. Assuming no significant multipath errors
obtained using GSNRx™ for PRN 28 and 17 for different were affecting that base station, differences between
antenna elements are shown in Fig. 10; PRN 28 is at high pseudoranges and true ranges provide combined meas-
elevation and PRN 17 at a low elevation. The rapid C/N0 urement errors as seen by the base station antenna.
variations of PRN 28 at all the antennas are comparable to Similarly, the approximate remote receiver position can
each other. PRN 17, which is affected by multipath, shows be obtained using the SPAN™ LCI unit with an accuracy
different C/N0 periodic variations, indicating reception of of few centimentres. Here, “remote receiver” refers to
different multipath signal phase values at different an- the referece antenna of the antenna array. Using the ap-
tenna elements. For PRN28, the mean C/N0 value is dif- proximate antenna position and ephemeris information,
ferent for different antennas. These differences are due to the true range can be obtained. By taking the difference
the gain patterns of different antenna elements and will be between pseudoranges and true ranges, combined meas-
corrected in the calibration process. urement errors as seen by the remote station can also be

50
PRN 17
C/N0 (dB-Hz)

45

40 Antenna-01
Antenna-02
35 Antenna-03
Antenna-04
Antenna-05
30
0 Antenna-06
50 100 150 200 250 300
50
Time (s) PRN 28
C/N0 (dB-Hz)

45

Antenna-01
40
Antenna-02
Antenna-03
35 Antenna-04
Antenna-05
30 Antenna-06
0 50 100 150 200 250 300
Time (s)
Fig. 10 C/N0 variations of PRN 17 and PRN 28 for different antenna array elements
Vagle et al. The Journal of Global Positioning Systems (2016) 14:4 Page 12 of 15

a 50

45

40 Before beamforming
DAS PRN 13
35
0 10 MPDR
20 30 40 50 60 70 80 90 100
60
C/N (dB-Hz)
MPDRSS
50

40
0

PRN 15
30
0 10 20 30 40 50 60 70 80 90 100
50

40

30
PRN 17
20
0 10 20 30 40 50 60 70 80 90 100
Time (s)
b 10
PRN 13
5

-5
0 10 20 30 40 50 60 70 80 90 100
5
PR Error (m)

PRN 15 Before beamforming

-5 DAS
0 10 20 30 40 50 60 70 80 90 100
50 MPDR
MPDRSS

PRN 17
-50
0 10 20 30 40 50 60 70 80 90 100
Time (s)
Fig. 11 a C/N0 and b pseudorange errors before and after beamforming

obtained. As the base station and remote receiver are PRNs removes it, finally yielding multipath errors. To
nearby, the differences between the pseudorange mea- perform double differencing, PRN 28, which is not af-
surements cancel out all the errors except multipath, fected significantly by multipath, was used as the refer-
user clock bias and measurement noise. Pseudorange ence satellite. The multipath errors for PRN 13,15 and
measurement noise was separately computed using a 17 before and after beamforming are shown in Fig. 11b.
zero-baseline and the standard deviation of the measure- Consider data between 80 and 100 seconds for analysis;
ment noise was measured as 8 cm for both GSNRx and
Novatel receivers. Therefore, by taking the differences Table 3 RMS pseudorange errors before and after beamforming
between base and remote receiver, the measurement PRN RMS pseudorange errors (m)
noise of pseudorange increases by 1.42 to 11 cm (Misra Before beamforming DAS MPDR MPDRSS
et al. 1996). However, compared to the magnitude of PRN 13 4.2 1.9 1.0 0.9
multipath errors at the metre level, it can be neglected PRN 15 2.1 1.3 1.3 1.8
for this evaluation. As the user clock bias is common for
PRN 17 8.9 0.9 1.2 1.2
all the PRNs, performing double differencing between
Vagle et al. The Journal of Global Positioning Systems (2016) 14:4 Page 13 of 15

40

East (m)
20

-20
0 20 40 60 80 100 120
North (m) 5

0
Before beamforming
-5 DAS
MPDR
-10
MPDRSS
20 0 20 40 60 80 100 120

0
Up (m)

-20
-40
-60
0 20 40 60 80 100 120
Time (s)
Fig. 12 Position errors before and after beamforming

the C/N0 degradation for antenna 1 (reference antenna) solutions computed using pseudorange measurements
is significant during this time interval and a similar deg- generated from the reference antenna are referred to as
radation can be observed with pseudorange errors. The the position solutions before beamforming. Similarly, pos-
beamformer is able to mitigate multipath and the RMS ition solutions computed using the pseudorange measure-
pseudorange error reduces from 20 m to 0.8 m after ments after beamforming are referred to as the position
beamforming using either of the three beamformers. solutions after beamforming. The reference position of the
The RMS pseudorange errors for PRN 17 considering antenna array was computed using the outputs of SPAN™
the entire data set reduces from 8.96 m to 0.92 m after LCI unit, which provides ultra-tight GNSS-INS solution
beamforming. The RMS pseudornage errors for different with accuracy of the order of at least a few decimetres.
PRNs for different beamforming techniques are shown Using the reference antenna array position, position errors
in Table 3. It can be observed that beamformer perform- before and after beamforming were computed and are
ance is different for different PRNs. This is due to differ- shown in Fig. 12. As only four satellites were visible, Pos-
ent satellite DOAs and multipath signals as mentioned ition Dilution of Precision (PDOP) is of the order of 10.
in the numerical simulation section. As shown after beamforming, the position errors are sig-
Performance of field test results are comparable with that nificantly reduced. The RMS position errors before and
of simulations. Considering PRN17 which is affected by after beamforming are provided in Table 4.
multipath, the performance of DAS beamformer depends
only on the LOS signal DOA. If the multipath signal direc- Conclusions
tions coincide with the beampattern nulls obtained from The numerical simulation results described in the paper in-
the DAS beamformer, it can provide comparable results to dicate that performance of MPDR and MPDRSS beamfor-
those of the MPDR beamformer, which is the case with mers improves as the correlation between LOS and
PRN 17. Similar performance of MPDR and MPDRSS are multipath signals decreases. It was observed that, for a rect-
likely due to the sufficient decorrelation between the LOS angular array with six antenna elements, the MPDRSS
and multipath signals over the 1 s integration considered to beamformer provides better multipath mitigation for higher
compute the covariance matrix. Also, as shown in Fig. 3, elevation satellites. The proposed multi-antenna signal
for lower elevation satellites with sufficient decorrelation
between LOS and multipath signals, the performance of Table 4 RMS position errors before and after beamforming
MPDR and MPDRSS beamformer are similar. RMS Position Errors East (m) North (m) Up (m)
The third analysis is performed to show the improve- Before beamforming 12.7 2.2 19.5
ment in position before and after beamforming. The least DAS 3.8 3.5 9.0
squares method was used to compute the position from MPDR 3.5 2.9 8.8
pseudorange measurements. Four observable satellites, MPDRSS 3.5 2.9 8.8
shown in green circles in Fig. 9b, were used. Position
Vagle et al. The Journal of Global Positioning Systems (2016) 14:4 Page 14 of 15

simulator was used to generate multipath affected multi- Arribas J, Closas P, Fernández-Prades C (2014) “Interference mitigation in GNSS
antenna signals for different user environments and the re- receivers by array signal processing: a software radio approach”, 2014 IEEE
8th Sensor Array and Multichannel Signal Processing Workshop (SAM), IEEE,
sults were compared with realistic multipath scenarios. A Coruna, p 121–124. doi:10.1109/SAM.2014.6882355
Using the simulated GPS signals, it was observed that pseu- Backén S, Akos DM, Nordenvaad ML (2008) “Post-processing dynamic GNSS
dorange errors can be reduced by tens of metres in high antenna array calibration and deterministic beamforming”. In Proceedings of
the 21st international technical meeting of the satellite division of the
multipath environments, thereby improving position accur- institute of navigation (ION GNSS 2008), vol 1619. Savannah, p 2806–2814
acy. It was observed that the MPDRSS beamformer per- Borre K, Akos DM, Bertelsen N, Rinder P, Jensen SH (2007) A software-defined GPS
forms better than the MPDR and DAS beamformer. With and galileo receiver: single-rrequency approach. Birkhäuser, Boston
Broumandan A, Jafarnia-Jahromi A, Daneshmand S, Lachapelle G (2016)
actual GPS L1 signals collected in a moderate specular mul- “Overview of spatial processing approaches for GNSS structural interference
tipath scenario, a reduction of 10 m in RMS pseudorange detection and mitigation”. In Proceedings of the IEEE, vol 99. p 1–12
error was observed for satellites affected by multipath sig- Brown A (2000) “Multipath rejection through spatial processing”. In Proceedings
of the proceedings of the 13th international technical meeting of the
nals. Pseudorange error reduction was reflected in the pos- satellite division of the institute of navigation (ION GPS 2000), Salt Lake City,
ition solutions. Finally, it was shown that a six-antenna p 2330–2337
rectangular array is effective to mitigate short-range multi- Chen YH, Juang JC, Seo J, Lo S, Akos DM, De Lorenzo DS, Enge P (2012) “Design
and implementation of real-time software radio for anti-interference GPS/
path signals and provide an improved navigation solution, WAAS sensors”. Sensors 12(10):13 417–13 440. doi:10.3390/s121013417
based on the data used in the analysis. Extensive testing Cuntz M, Konovaltsev A, Meurer M (2016) “Concepts, development, and
would be required to confirm these enhancements in differ- validation of multiantenna GNSS receivers for resilient navigation”. In
Proceedings of the IEEE, vol 99. p 1–14
ent environments.
Daneshmand S, Broumandan A, Nielsen J, Lachapelle G (2013a) “Interference and
multipath mitigation utilizing a two-stage beamformer for global navigation
Authors’ contributions
satellite systems”. In IET, Radar, Sonar and Navigation 7(1):55–66
NV involved in the major contributions such as literature review, software
simulator and receiver development, data collection and processing the data Daneshmand S, Broumandan A, Sokhandan N, Lachapelle G (2013b) “GNSS
and preparing the manuscript. AB helped in developing software simulator multipath mitigation with a moving antenna array”. IEEE Trans Aerospace
and data collection. AJ participated in technical discussions regarding Electro Syst, 49(1):693–698
numerical simulations and software simulator development. GL participated Daneshmand S, Jafarnia-Jahromi A, Broumandan A, Lachapelle G (2014) “A GNSS
in discussing the methodology and the live data results. All authors read and structural interference mitigation technique using antenna array processing”.
approved the final manuscript. In 2014 IEEE 8th Sensor Array and Multichannel Signal Processing Workshop
(SAM), A Coruna, p 109–112
Authors’ information Dierendonck AJV (1996) “GPS Receivers”. In: Parkinson B and Spilker JJ Jr. (eds)
Niranjana Vagle is a Ph.D. candidate in the PLAN Group of the University of [Chapter 8] in global positioning system: theory and applications, vol 1.
Calgary. He has 5 years of industry experience in GPS-GLONASS receiver hardware American Institute of Aeronautics and Astronautics, Inc., Washington D.C., p
design, baseband signal processing, and environment acceptance tests. He has 329-408
received his B.E. in Electronics and Communications Engineering in 2005 from Dierendonck AJV, Fenton P, Ford T (1992) “Theory and performance of narrow
Visvesvaraya Technological University (VTU), India. His main research interest is correlator spacing in a GPS receiver”. J Inst Navigation 39(3):265–283
antenna array signal processing for GNSS applications. Egea D, López-Salcedo JA, Seco-Granados G (2014) “Interference and multipath
Dr. Ali Broumandan received his Ph.D. degree in the Geomatics Engineering sequential tests for signal integrity in multi-antenna GNSS receivers”. In 2014
from the University of Calgary. He is working in the PLAN Group as Senior IEEE 8th Sensor Array and Multichannel Signal Processing Workshop (SAM), A
Research Associate since November 2013 where his research focuses on Coruna, p 117–120
GNSS interference mitigation utilizing single and multiple antenna Fern’andez-Prades C, Closas P, Arribas J (2011) “Eigenbeamforming for interference
processing. He has been involved in several industrial research projects mitigation in GNSS receivers”. In Proceedings of the 1st International
focusing on spatial/temporal GNSS signal processing. Conference on Localization and GNSS (ICLGNSS’11), p 93–97
Dr. Ali Jafarnia Jahromi received his Ph.D. in Geomatics Engineering from the Fern’andez-Prades C, Arribas J, Closas P (2016) “Robust GNSS receivers by array
University of Calgary. He holds B.Sc. and M.Sc. degrees in signal processing: theory and implementation”, In Proceedings of the IEEE,
Telecommunications Engineering. He was working as Post-Doctoral Fellow in vol 99. p 1-14
the PLAN Group from 2013 to 2016. His research interests include signal Fu Z, Hornbostel A, Hammesfahr J, Konovaltsev A (2003) Suppression of
processing in GNSS applications and receiver design. multipath and jamming signals by digital beamforming for GPS/Galileo
Gérard Lachapelle, Professor Emeritus, has been involved in a multitude of applications. GPS Solutions 6(4):257–264
GNSS R&D projects since 1980, ranging from RTK positioning to indoor Garlin L, VanDigelen F, Rousseau JM (1996) “Strobe and edge correlaor multipath
location and signal processing enhancements, first in industry and since mitigation for code”. In Proceedings of the proceedings of the 9th
1988, at the University of Calgary. international technical meeting of the satellite dicision of the institute of
navigation, Kansas City, p 657–664
Competing interests Gupta IJ, Weiss IM, Morrison AW (2016) “Desired features of adaptive antenna
The authors declare that they have no competing interests. arrays for GNSS receivers”. In Proceedings of the IEEE, vol 99. p 1-12
Joosten P, Pany T, Winkel J (2002) The impact of unmodelled multipath on ambiguity
Received: 28 May 2016 Accepted: 7 September 2016 resolution. In: Proceedings of the proceedings of the 15th international technical
meeting of the satellite division of the institute of navigation (ION GPS 2002).
Oregon Convention Center, Portland, pp 953–961
References Kalyanaraman SK, Braasch MS (2007) “Tight Integration of a GPS adaptive array
Amin MG, Wang X, Zhang YD, Ahmad F, Aboutanios E (2016) “Sparse arrays and with a software-defined receiver”. In Proceedings of the 2007 national
sampling for interference mitigation and DOA estimation in GNSS,” in technical meeting of the institute of navigation, San Diego, p 657–668
Proceedings of the IEEE, vol 99. p 1–16 Kalyanaraman SK, Braasch MS (2010) “GPS adaptive array phase
Arribas J, Closas P, Fern’andez-Prades C, Cuntz M, Konovaltsev A, Meurer M (2012) compensation using a software radio architecture”, Navigation, J Inst
“Advances in the theory and implementation of GNSS antenna array Navigation 57(1):53-68
receivers. In: Georgiadis A, Rogier H, Roselli L, Arcioni P (eds) Microwave and Konovaltsev A, Antreich F, Hornbostel A (2007) “Performance assessment of
millimeter wave circuits and systems: emerging design, technologies and antenna array algorithms for multipath and interference mitigation” in Proc.
applications. Wiley, Chichester, pp 227–273 2nd Workshop GNSS Signals & Signal Process, ESTEC, Noordwijk
Vagle et al. The Journal of Global Positioning Systems (2016) 14:4 Page 15 of 15

Lee J-H, Hsiao S-W (2008) “adaptive processing for beamforming with coherent
interference using circular arrays”. Digital Signal Process 18(5):813–834
Manosas-Caballu M, Seco-Granados G, Swindlehurst AL (2013) “robust
beamforming via FIR filtering for GNSS multipath mitigation” in Proceedings
of the 2013 IEEE International Conference on Acoustics, Speech and Signal
Processing (ICASSP), Vancouver, p 4173–4177
Misra P and Enge P (1996) Global Positioning System, Signals, Measurements and
Performance, vol. 2. Ganga-Jamuna Press, ISBN 0-9709544-1-7
Novatel GPSAntennaTM model 501 (http://www.novatel.com/assets/Documents/
Manuals/om-20000001.pdf). Accessed 10 Apr 2016
Petovello MG, O’Driscoll C, Lachapelle G, Borio D, Murtaza H (2008) architecture
and benefits of an advanced GNSS software receiver. J Global Positioning Sys
7(2):156–168
Ray JK, Cannon ME, Fenton P (1999) “Code Range and Carrier Phase Multipath
Mitigation Using SNR, Range and Phase Measurements in a Multi-Antenna
System”, Proceedings of ION GPS-99 (Nashville, September 14–17)
Reddy V, Paulraj A, Kailath T (1987) “analysis of the optimum beamformer in the
presence of correlated sources and its behavior under spatial smoothing”.
IEEE Trans Acc Signal Process 35(7):927–936
Rougerie S, Carrie G, Ries L (2011) “Multipath mitigation methods based on antenna
array,” in Proceedings of the International Technical Meeting of Institute of
Navigation (ITM’11), San Diego, p 596–605
Rougerie S, Vincent CGF, Ries L, Monnerat M (2012) “A new multipath mitigation
method for GNSS receivers based on antenna array,” Intl J Navigation
Observation 2012:804732. doi:10.1155/2012/804732
Sahmoudi M, Amin M (2007), “Optimal robust beamforming for interference and
multipath mitigation in GNSS arrays,” in Proc. IEEE int. conf. acoust. speech
signal process, vol. 3. Honolulu, p 693–696
Seco-Granados G, Fernandez-Rubio JA, Fernandez-Prades C (2005) “ML estimator
and hybrid beamformer for multipath and interference mitigation in GNSS
receivers”. IEEE Trans Signal Process 53(3):1194–1208
Van Nee R, Siereceld J, Fenton PC, Townsend BR (1994) “Multipath estimating
delay lock loop: approaching theoretical accuracy limits”, in Proceedings of
the position location and navigation synopsium, IEEE Las Vegas, Nevada, p
246–251
Van Trees HL (2002) Optimum array processing, detection, estimation, and
modulation theory part IV. Wiley, New York, pp 600–699
Vicario JL, Barcelo M, Mañosas M, Seco-Granados G, Antreich F, Cebrian JM, Picanyol
J, Amarillo F (2010) “A novel look into digital beamforming techniques for
multipath and interference mitigation in Galileo ground stations”, In: 2010 5th
advanced satellite multimedia systems conference and the 11th signal
processing for space communications workshop, p 240–247
Widrow B, Duvall K, Gooch RP, Newman W (1982) “Signal cacellation phenomena in
adaptive antennas: causes and cures”. IEEE Trans Antennas Propag 30(3):469–478

Submit your manuscript to a

journal and benefit from:
7 Convenient online submission
7 Rigorous peer review
7 Immediate publication on acceptance
7 Open access: articles freely available online
7 High visibility within the field
7 Retaining the copyright to your article

Submit your next manuscript at 7 springeropen.com