A Novel Approach in Detection and Characterization of CW

Document Sample
A Novel Approach in Detection and Characterization of CW Powered By Docstoc
					    A Novel Approach in Detection and
Characterization of CW Interference of GPS
 Signal Using Receiver Estimation of C/No
                           A. T. Balaei, A.G. Dempster and J. Barnes
               Cooperative Research Centre for Spatial Information System (CRC-SI)
                  and the School of Surveying and Spatial Information Systems at
                               The University of New South Wales

ABSTRACT            Narrowband interference can       these primary modules, it can also affect the
severely degrade the performance of GPS               carrier and code tracking loops [3, 4] which
receivers. Different narrowband interference          results in deterioration of all the GPS
suppression techniques also degrade this              observables or in complete loss of lock in severe
performance correspondingly. Detecting the            cases. Continuous-wave (CW) interference has
presence of interference and then characterizing      been proved to have effects for the GPS C/A
it can lead to its removal. Knowledge that can be     code signal [5], related to the characteristics of
useful is: the location of the source or direction    the frequency spectrum of this signal. To make
of arrival of the interference (spatial-directional   the receiver less vulnerable to interference or to
characteristics), the time specification and          monitor the quality of the signal, interference can
frequency and power of the interference               be detected in the receiver either before the
(temporal-spectral characteristics). Our focus in     correlation (pre-correlation detection) or after it
this work is on the second type. Using the post-      (post-correlation detection). In the pre-
processing capability of a software GPS receiver,     correlation techniques, antenna [6], AGC [2] and
CW interference is detected and characterized.        ADC [7] have been used to detect and
This is achieved by passing the GPS signal and        characterize the RFI and in post-correlation
the interference through the correlator. After        techniques, the observables of the receiver that
correlation, using the definition of carrier to       are affected by RFI have usually been used to
noise density ratio (C/No), a mathematical            detect and characterize the interference. In [8],
expression for C/No is given in which the             using a statistical approach, it has been shown
temporal and spectral parameters of interference      that correlator output power shows consistent
are found. Then, using the conventional               performance under varying levels as well as
definition of C/No as the squared mean of the         types of interference and carrier phase vacillation
correlator output divided by its variance, the        was used as a backup indicator. In [9], correlator
actual C/No is calculated. Finally comparing          output is calculated in a multi-correlator receiver
these two values and considering the structure of     to estimate the frequency of the CW RFI. In this
the GPS C/A code, the existence and                   paper, AGC level together with correlator output
characterization of the existing interference is      power is used to detect and characterize the RFI
determined.                                           but the difference is that the receiver simply uses
                                                      a standard correlator. The other advantage of this
KEYWORDS: GPS,             Interference,    C/No,     algorithm is that it is capable of detecting and
Correlator                                            characterizing CW RFI even it is not close in
                                                      frequency to any of the C/A code spectral lines.
              I. INTRODUCTION                            The problem is described in section 2. In
  Radio frequency interference (RFI) is amongst       section 3, a mathematical expression for the
the most disruptive events in the operation of a      C/No which is introduced. We derived it in [4],
GPS receiver. It affects the operation of the         using the spectral analysis of both the signal and
automatic gain control (AGC) and low noise            the RFI in passing through the correlator. The
amplifier (LNA) in the RF front-end [1, 2] and        actual C/No is calculated using received I and Q
depending on how much of it passes through            data from a software GPS receiver in section 4.
Hardware setup for the experiments is presented
in section 5. The spectral parameters of RFI are
compared in section 6 and finally section 7
concludes this paper.
                                                      Figure 4 Tracking loop low pass filter, filters the
           II. PROBLEM DEFINITION                       data and the interference which is outside the
  The GPS C/A code is a Gold code with a                             filter bandwidth out
relatively short 1-ms period (i.e., the PRN
sequence repeats every 1 ms). Therefore, the C/A          As the code is despread by getting multiplied
code (neglecting the navigation data) has a line      by the replica code in the receiver, interference is
spectrum with lines 1 kHz apart [1]. In Figure 1,     spread over the frequency bandwidth of the
navigation data is incorporated into the code and     original signal in the same way that the data is
in Figure 2, noise and interference is added to the   spread in the satellite (Figure 3). A low pass
signal to achieve the final GPS signal received at    filter is used in the tracking loops. Only
the antenna.                                          interference that is within the bandwidth of this
                                                      filter remains (Figure 4). In figure 2, the
                                                      amplitude of the interference is shown to be
                                                      Jbefore. In passing the RFI through the correlator
                                                      and the filter, the value of the amplitude of the
                                                      remaining interfering signal is Jafter. The aim of
 Figure 1 Spreading the data over the C/A code        these figures is to show that the value Jafter is
             spectrum bandwidth                       determined by the strength of the nearest line to
                                                      the interference (in this example jth line). Now if
                                                      we have an RFI with fixed frequency and a GPS
                                                      signal with Doppler frequency that varies with
                                                      satellite motion, over time the RFI coincides
                                                      with several different consecutive lines in the
                                                      spectrum. Each of these lines has its own unique
 Figure 2 Interference and background noise is        effect on the remaining interference in the output
 added to make the final GPS signal received at       of the loops. We propose to examine the effect of
                  the antenna                         a series of lines, and thus calculate the frequency
                                                      of the RFI. The quantity that can best reflect this
Interference is assumed to be CW constant             effect is the correlator output power. Carrier to
                                                      noise ratio which quantifies the quality of the
amplitude (CA) [10] at the frequency f i away         signal is the parameter which is used for this
from the band center of the code spectrum and         purpose. In the following section, this value is
 ∆f i away from the nearest Dirac line in the         parametrically drawn in the presence of CW RFI.
spectrum. This line is also assumed to be the jth
line of the spectrum. In other words this line is j
kHz away from the band center as all the lines                   III. C/NO MATHEMATICAL
are 1 kHz apart from each other. Now in the                               EXPRESSION
base-band processing of the GPS receiver, this
signal is passed through the correlator to be            Many receivers report received signal quality
acquired and tracked. In the correlator, first the    as carrier-power-to-noise density ratio, denoted
code is despread (Figure 3) and then the              C/No. In [3], this receiver report has been
navigation data is filtered out (Figure 4).           characterized in terms of front-end bandwidth,
                                                      discriminator design and other receiver
                                                      characteristics. This characterization is done
                                                      using the conventional definition of C/No as the
                                                      squared mean of the correlator output divided by
                                                      its variance. In [4], as a special case, assuming
                                                      the interference to be CW constant amplitude
Figure 3 Code is despread by getting multiplied
                                                      signal, another expression is derived for C/No
 by the receiver code replica and interference is
                                                      which is consistent with the previous results and
                                                      in which spectral parameters of the RFI can be
found. Figure 5 shows the final result of Figure 2                      spectrum. The other point which is noticeable in
as the received GPS signal passing through the                          this graph is the sinc functions occurring around
correlator.                                                             each trough. The width of each sinc function is
                                   Code Tracking Loop
                                                                        related to the integration period, as can be seen
                                                                        in the Eq. 1. The longer the integration period is
                                                                        the narrower will be the sinc functions.
                                       Carrier Tracking Loop
                    CW                 I                       Output

                                                                             C/No (dBm)
                                       OSC              LPF

                          90                                   arctan                     36

                                       Q                                                  34

Figure 5 Correlator (Code and Carrier Tracking
                    Loops)                                                                30

                                                                                               0       1000   2000   3000     4000     5000 6000    7000   8000   9000

Eq. 1 shows the mathematical expression for the                                                                             Doppler Frequency

C/No in the output of the correlator.                                          Figure 6 C/N0 calculated using the
                                                                          mathematical expression for satellite 1 with
                                                                         Doppler frequency changing from 0 kHz to 10
             (Td R0 (τ ).sinc(∆f cTd )) 2
C/No =                                            Eq. 1                  kHz and CW interference at 14 kHz away from
          Ln N 0 + J (Td .C j .sinc(Td .∆f i )) 2                              the band center at 1.57542 GHz.

         N0 is the thermal noise power                                       IV. ACTUAL CALCULATION OF C/No
         Ln is the processing gain in the noise                            There are different techniques for estimating
         Td is the integration duration time                            the carrier power to noise density, C/N0. This
         τ is the signal-reference code phase                           estimate is important because it helps determine
difference in code chips                                                whether the code and carrier tracking loops are
           ˆ                                                            in lock, controlling the response of the receiver
           f c is the estimate of carrier frequency.                    to low signal to noise environments, and
           ∆f = f - fˆ
                c     c        c
                                                                        determining the signal to noise environment in
                                                                        order to assess or predict receiver performance.
           ∆f i = f i - f c                                             In [11] is a review of different measurement
          J is the interference power                                   techniques such as the narrow-to-wideband
          Cj is the jth spectral line coefficient                       power ratio method, correlator comparison
           R0 (τ ) is the cross correlation of the                      method, unnormalized discriminator output
                                                                        statistics method. It was found in [11] that the
received C/A code and the receiver replica of the                       power ratio method performs better in terms of
same code.                                                              noise in low signal-to-noise environments. This
   In Figure 6, as an example, using Eq. 1 and                          method is widely known and in [12], is
assuming a specific environmental noise power,                          presented. The prompt I and Q samples over the
the C/No is drawn for satellite 1 with Doppler                          accumulation intervalτ , are divided into M
frequency changing from 0 kHz to 10 kHz and                             intervals. These samples are then used to
CW interference at 100 kHz away from the band                           calculate a narrowband power, PN, over the
center at 1.57542 GHz.                                                  whole accumulation interval and a wideband
   The deep troughs in this graph correspond to
                                                                        power, PW, over the interval τ / M , then
the coincidence of CW RFI with the code
spectral lines. It is clear from the picture that this                  summed over τ . These power estimates are:
happens at 1 kHz spacing in the Doppler. As
                                                                                               M                        M
expected and explained in the previous section,
there are different values for different lines. This                    PN = (∑ I Pi ) 2 + (∑ QPi ) 2                                              Eq. 2
                                                                                                   i                        i
difference comes from the difference between
the coefficients of different lines in the code
PW = (∑ ( I Pi + QPi ))
            2     2
                                       Eq. 3

where          I Pi = 2(c / n0 )τ / M cos ϕ + wIPi
and Q Pi =        2(c / n0 )τ / M sin ϕ + wQPi
where wIP and wQP are normalized random noise
samples from a zero mean unit variance normal
(Gaussian) distribution. The narrow-to-wide
power ratio, PN/W, is simply the ratio of the two
power measurements. However, to reduce the
noise, the measurement is averaged over n
iterations. Thus,
                                                         Figure 7 Hardware setup for the experiments
               1 k PN ,r
PN / M =         ∑
               k r =1 PW ,r
                                       Eq. 4               from left: Spectrum analyzer, RF signal
                                                          generator, NordNav front-end, GPS signal
         In [12], it is shown how to derive      Eq.                      generator.
5 from the above equations.
                   M (c / n0τ + 1)                                            55

E ( PN / W ) ≈                         Eq. 5
                    M + c / n0τ                                               50

where E() is the expectation operator.                    Actual C/No (dBm)   45

Rearranging this gives the measured carrier                                   40
power-to-noise density as a function of the
power ratio measurement:                                                      35

           M PN / W − 1
c / n0 ≈                Eq. 6                                                 30

           τ M − PN / W
                                                                                   0   200   400   600   800    1000   1200   1400   1600   1800
                                                                                                         Time (Sec)
  In [12], it is shown graphically that for
                                                        Figure 8 C/N0 calculated using the power ratio
c/n0>23dB there is less than 1dB estimation error
                                                       technique for satellite 1 with Doppler frequency
for an average time of 1 s (M=20 and k=50). For
                                                        changing from 0 to 9 kHz and CW interference
smaller values of C/N0 longer averaging time is
                                                       at 14 kHz away from the band center at 1.57542
          V. HARDWARE SETUP FOR                           There are a number of points regarding this
               EXPERIMENTAL RESULTS                    figure which is addressed as follows. At the
   In Figure 7, the hardware setup to measure the      beginning, it has a large value (50 dB), which is
actual C/N0 is shown. The NordNav [13]                 just an initial condition and gradually converges
software receiver is used to capture the IF data to    to the real actual value of C/N0. Figure 9 and
be analyzed and post processed and an HP8648B          Figure 10 are focused on one of the troughs
is used to generate the CW interference which is       where RFI coincides with a code spectral line in
combined with the GPS signal generated by a            the actual measurement and theoretical
SPIRENT GSS6560. In Figure 8, the actual               calculation of C/N0 correspondingly. The first
measured C/N0 using NordNav software GPS               noticeable thing is the small peak in the actual
receiver is illustrated.                               estimation right in the middle of the trough of
                                                       the sinc function. We have explained this
                                                       phenomenon in [4] using a DataFusion [14]
                                                       software receiver. Around this peak, as
                                                       interference frequency is very close to the carrier
                                                       frequency, it is actually helping the C/N0. This
                                                       phenomenon can not be helpful in our proposed
                                                       detection-characterization method. By choosing
appropriate M and K in the power ratio
estimation of C/N0 (        Eq. 5), this peak can
be averaged out. The only disadvantage of                                                                45

choosing K and M such that the peak is averaged                                                          40
out is that some of the information in the C/N0 is

                                                                                     Actual C/No (dBm)
also lost. This information fortunately is not                                                           35

essential to this method as it is only the peak
information of the lines which are used.

                     50                                                                                  25


    c a /N B )
   A tu l C o (d m


                     40                                                                                           100     200     300    400     500    600   700     800    900    1000
                                                                                                                                               Time (sec)


                                                                                  Figure 11 C/N0 calculated using the power ratio
                                                                                  technique for satellite 1 with Doppler frequency

                                 1300     1350            1400     1450    1500

                                                                                       changing from -4 kHz to 4 kHz and CW
                                                 Time (s econd)

                                                                                   interference at the band center at 1.57542 GHz
     Figure 9 One of the troughs where RFI
 coincides with a code spectral line in the actual
                                                                                     In Figure 12, C/N0 plots calculated using both
                                                                                  the parametric method and power ratio method,
                                                                                  are drawn together. The fact that for both
                                                                                  techniques, the relationship between the values
                                                                                  of any two consecutive peaks remains the same
     /N B )
    C o(d m

                                                                                  is illustrated. In Figure 12, in fact we have
                                                                                  Figure 6 and Figure 8 which is flipped over and
                                                                                  the horizontal axis is converted to Doppler
                                                                                  frequency. The reason why it should be flopped
                          1400   1600   1800     2000      2200
                                          Doppler Frequency (Hz)
                                                                   2400   2600
                                                                                  over is that in an increasing Doppler frequency
                                                                                  scenario, as interference cross the spectral lines,
          Figure 10 One of the troughs where                                      the last line which is crossed is in fact the lowest
  RFI coincides with a code spectral line in the                                  frequency spectral line.
          theoretical measurement                                                                        55
                                                                                                                                                                     Actual C/No
                                                                                                                                                                     Theroretical C/No
      VI. CHARACTERIZING THE CW RFI                                                                      50

  The results in sections IV and III are compared
in this section to extract the spectral information
                                                                                     C/No (dBm)

of the CW RFI. As it was explained the peak                                                              40
information is used. The fact that each C/A code
has a unique pattern of spectrum is used in this                                                         35

comparison. In Figure 11, the uniqueness of this
pattern is shown. The Doppler frequency of                                                               30

satellite 1 is changed from -4 kHz to 4 kHz. And
the CW RFI is placed right in the band center.                                                                0    1000         2000    3000     4000    5000
                                                                                                                                        Doppler Frequency (Hz)
                                                                                                                                                                    6000     7000        8000

The symmetry in the picture can clearly show
that coincidence of interference with a line
                                                                                   Figure 12 C/N0 plots calculated using both the
always results in the same pattern of C/N0.
                                                                                    parametric method and power ratio method

                                                                                     In Figure 13, the difference between the
                                                                                  estimated C/N0 and the theoretically calculated
                                                                                  one is shown. This difference has been
                                                                                  calculated in the first 1000 kHz from the band
                                                                                  center. This is in fact the difference between the
                                                                                  8 trough values of Figure 8 and the values of 8
                                                                                  troughs of an “8 trough searching window” over
the whole bandwidth of the C/A code spectrum.                                                                      In this work, the power of interference is
The minimum difference in this experiment is at                                                                  estimated by the value of AGC, which works
14 kHz which is the frequency at which the RFI                                                                   relative to the background noise power. So if the
is added to the GPS signal. For clear vision, only                                                               difference between the assumed and real
the first 80 kHz is shown. In the Eq. 1, other than                                                              background noise power very different, then the
the frequency of interference, there are three                                                                   result of this technique wouldn’t be reliable. To
other parameters: Signal power, background                                                                       address this problem, we can calculate the
noise power and interference power. The first                                                                    environmental noise power by acquiring one of
two are known parameters as the GPS signal has                                                                   the satellite signals.
been generated in a known environment
condition and with a desired signal power.                                                                       REFERENCES
Interference power is also estimated from the
AGC level in the RF front-end [2]. Using this                                                                    [1]    Kaplan E.,   (1996) “Understanding
value we can have the theoretical C/No fit the                                                                   GPS: Principles and Applications”, Artech
actual C/No as much as possible. This will allow                                                                 House.
us having less difference between the two
calculations at the frequency where interference                                                                 [2]      F.  Bastide,     C.      Macabiau,
exists (14 kHz in this experiment).                                                                              DM Akos “Automatic Gain Control (AGC) as an
                                                                                                                 Interference  Assessment     Tool”,    ION
                                                                                                                 GPS/GNSS, (2003)
  Difference between values of 8 consecuitive troughs (dBm)


                                                              90                                                 [3]      J. W. Betz, “Effect of Partial-Band
                                                              80                                                 Interference on Receiver Estimation of C/N0:
                                                              70                                                 Theory,” Proceedings of ION 2001 National
                                                              60                                                 Technical Meeting, Institute of Navigation,
                                                              50                                                 January 2001.

                                                              30                                                 [4]      A. T. Balaei, J. Barnes, A. G. Dempster,
                                                                                                                 “Characterization of interference effects on GPS
                                                                                                                 signal carrier phase error” SSC (2005)
                                                                    10   20   30         40       50   60   70
                                                                                   Frequency (kHz)
                                                                                                                 [5]      Spilker J. and Natali F., (1996)
                                                                                                                 “Interference    Effects    and     Mitigation
      Figure 13 The difference between the                                                                       Techniques”, Chapter 20 of ‘Global Positioning
 estimated C/N0 and the theoretically calculated                                                                 System: Theory and Applications’, AIAA.
                                                                                                                 [6]     Alison Brown, Sheryl Atterberg, and
     VII. SUMMARY AND FUTURE WORK                                                                                Neil Gerein, NAVSYS Corporation, “Detection
   In this paper, a new technique for detection                                                                  And Location Of GPS Interference Sources
and characterization of CW interference was                                                                      Using Digital Receiver Electronics” Proceedings
presented. The advantage of this technique is that                                                               of ION Annual Meeting, June 2000, San Diego,
it uses the post processing capability of GPS                                                                    CA
software receivers without any need for extra
hardware. To do this, the spectral effect of CW                                                                  [7]     Frank Amoroso, “ Adaptive A/D
RFI on the C/A code is analyzed and an                                                                           Converter to Suppress CW Interference in DSPN
expression for C/N0 as a measure of this effect is                                                               Speard-Spectrum     Communications”       IEEE
parametrically drawn. Also using a technique for                                                                 Transaction on Communications, October 1983
C/N0 estimation, and the post processing
capability of a software GPS receiver, carrier                                                                   [8]     Awele Ndili, Per Enge, “GPS Receiver
power to noise ratio was calculated for a piece of                                                               Autonomous Interference Detection” Presented
pre-generated data containing a GPS signal of                                                                    at the 1998 IEEE Position, Location and
one particular satellite with a wide range of                                                                    Navigation Symposium - PLANS ‘98
variation in Doppler frequency. Finally
comparing these two values, the frequency of the                                                                 [9]    F. Bastide, E. Chatre, STNA, France; C.
CW interference is calculated.                                                                                   Macabiau, ENAC, France “Gps Interference
Detection      And      Identification  Using
Multicorrelator Receivers” ION GPS 2001

[10]     Moelker D., “Interference in satellite
navigation and mobile communication” Delft
University Press, Mekelweg 4, 2628 CD Delft,
Netherlands, 1998

[11]      Paul D. Groves, “ GPS Signal-to-Noise
Measurement in Weak Signal and High-
Interference Environments” Journal of the
Institute of Navigation. Vol. 52, No. 2, Summer

[12]     Van Dierendonck, A. J., GPS receivers,
In: B. W. Parkinson and J. J. Spilker (Eds.),
Global Positioning System: Theory and
applications, Volume I, AIAA, Washington, DC,
1996, pp. 329-408

[13]    DataFusion Corporation (2004) Kent
Krumvieda, GPS Program Manager Data Fusion
Corporation 10190 Bannock Street Suite 246
Northglenn, CO 80260 USA Tel: 720-872-2145
X303 Fax: 720-872-6418

[14] NordNav           Technologies       AB,
Stadsgården 10, S-116 45 Stockholm, Sweden,
Phone: +46 8 390 000, Fax +46 8 412 47 40

Shared By:
Description: A Novel Approach in Detection and Characterization of CW ...