22nd AIAA International Communications Satellite Systems Conference & Exhibit 2004 AIAA 2004-3182

9 - 12 May 2004, Monterey, California

GPS CDMA NOISE ANALYSIS

Dr. Srini Raghavan, Dr. Kuang Tsai and Mr. Lamont Cooper
The Aerospace Corporation

Abstract life applications, which implies that more stringent

availability requirements will be necessary. A number
Spectral lines of the GPS C/A codes play an important of augmentation systems both ground and land based
role in the amount of Code Division Multiple Access are planned and being implemented, such as WAAS
(CDMA) noise generated, which degrades the available (Wide Area Augmentation System) and LAAS (Local
carrier-to-noise density ratio (C/N0) to the GPS Area Augmentation System), which use C/A codes for
receiver. This in turn affects the code acquisition and spreading the signal. A number of GPS-like satellite
tracking performance of the GPS signal, which are the navigation systems such as the European Galileo—
two operations that must be performed in a GPS which plans to use the same frequency band as the GPS
Downloaded by BEIHANG UNIVERSITY on April 3, 2020 | http://arc.aiaa.org | DOI: 10.2514/6.2004-3182

receiver before a complete navigation solution is [1] but may or may not use C/A codes—are in
calculated. Most analysis performed to date ignores the developmental stages. Newer blocks of the GPS are
spectral line effects altogether, claiming that any designed that also use C/A codes for the purpose of
significant effect is of short duration and limited providing backward compatibility to the existing GPS
geographical extent. The Aerospace Corporation felt receivers. So with the increase of so many C/A code
otherwise, that the spectral line effects couldn’t be transmitters, it is extremely important to accurately
ignored, especially with GPS band sharing with other calculate the code division multiple access (CDMA)
systems providing unlike-kind services. Furthermore, noise generated by these codes to make sure that there
certain GPS applications are categorized as safety-of- is no surprise outage of the service in a critical time of
life types of services, making it even more important to need. The problem is compounded further, although
be rigorous in the CDMA noise calculations. A rigorous only in limited geographic regions and limited periods
analysis is not only complex to perform but also time of time, in that the CA code signal has spectral lines
consuming to consider each and every scenario as it spaced 1 kHz apart because of the short period of the
arises in the future. To overcome this difficulty, a two- code, which may result occasionally in much higher
step procedure is proposed in this paper for performing CDMA noise than from a long code with no spectral
the interference analysis using the C/A code. lines. Since it is much harder to accurately quantify the
spectral line effects, past interference analyses
performed by various people resorted to some kind of
average performance while ignoring the worst-case
Introduction spectral line effects altogether.

In order to keep the code acquisition time to a In an effort to accurately quantify the worst-case C/A
reasonable value, Gold codes of length 1023 chips, at a code CDMA noise, The Aerospace Corporation has
chipping rate of 1.023 mega chips per second (MCPS) developed equations to accurately calculate the C/A
were chosen for use as coarse acquisition (C/A) codes code CDMA noise both in the code tracking and the
in the Global Positioning System (GPS). Gold codes acquisition modes. The analysis has shown that the
also had other desirable features such as, well-defined CDMA noise depends on received power levels,
auto- and cross-correlation properties and a large receiver antenna gain towards the interference signals,
number of codes in the code set available for use by the Doppler frequency shifts of the received signal, path
GPS. Selection of this code was made in the early ’70s, delays, and to a lesser degree on the particular C/A
based on the processor technology available then. For a code or codes of the interfering signal. In other words,
reasonable number of satellites using this code, the the user-satellite geometry plays a very significant role
code choice was quite good in taking advantage of all in the amount of CDMA noise experienced by the GPS
the desired properties of the Gold codes without being receivers. A software tool called CLIMAT (CDMA
hindered by any of its limitations. The situation has Limited Interference Modeling and Analysis Tool)—
changed significantly since then in a number of ways. which takes into account the user-GPS satellite’s
The user community is not just military, although the geometry, received power level variations, antenna gain
system was developed for the military use. The civilian patterns, and the C/A code spectral line effects—has
use of GPS has grown in proportions nobody been developed by the Aerospace Corporation.
anticipated, and some of those uses include safety-of-

1
American Institute of Aeronautics and Astronautics
Copyright © 2004 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved.
 C/A Code CDMA Noise Calculation

The acquisition mode is the most sensitive of all of the Since the above-mentioned effects are a function of
modes in a GPS receiver. Therefore, this study used the satellite and receiver geometry, a GPS constellation-
signal-to-noise-density (C/N0) level necessary for driven model of the satellite orbits and antenna
acquisition of the C/A code as the criterion for specification is used to model all the effects of
interference to GPS. Cross-correlation properties of the interference. This is particularly necessary for short-
C/A code (CDMA noise) make it more susceptible to duration effects on the order of minutes, since safety-
interference from increased signal levels of another C/A of-life uses of GPS have very stringent integrity,
code. This is why the analysis of the possible use of availability, and continuity requirements. The use of
PRN codes in the GPS C/A (Gold) code family by other the CDMA noise model, coupled with the geometry and
systems is particularly important. antenna effects, is recommended to handle interference
analysis for a Global Navigation Satellite System
Although it has been widely known that the C/A codes (GNSS).
used for GPS have a limited CDMA capacity [2], due in
Downloaded by BEIHANG UNIVERSITY on April 3, 2020 | http://arc.aiaa.org | DOI: 10.2514/6.2004-3182

part to their short (1-millisecond) length, earlier studies Using this methodology, a CDMA analysis tool termed
[3] that focused on interference to the system used the Aerospace-developed CDMA Limited Interference
Gaussian noise models for the C/A code CDMA noise Model and Analysis Tool (CLIMAT) [2] was
from the GPS. This simplification, in which the cross- developed. A block diagram model is shown in Figure
correlation effects are random, was made despite the 2. Note that the CDMA noise density I0 in Figure 2 is
fact that the mutual interference of the C/A codes in a not a constant unlike that typically used in other radio
worst-case scenario is much greater than that from frequency systems (such as communications), where
random noise. Therefore, when analyzing the only average interference effects are of interest (CDMA
performance of GPS in the presence of signals from noise density in this case is a constant, which is
other sources (depending on the GPS user and the inversely proportional to the spreading-code chipping-
specific requirements as discussed below), a more rate [3] due to the relative randomness among the
accurate model of CDMA noise from the GPS should received code phases.).
be used. Although a full description of the preferred
model of the C/A code as a short code is planned to be RECEIVED
SIGNALS FROM User AUGMEN-
in a later report, it is sufficient to note here that the THE GPS Location TATIONS
results of this model are a function of the differential CONSTELLATION

ranges and Doppler frequencies of the affected CA,P,M,IM

USER OTHER
HPM GNSS
satellites. In addition to the effects of the GPS signals, ANTENNA
CONSTELLATIONS
MODEL
the way in which GPS is used plays a large role in how
interference should be analyzed. Different GPS signals INTERFERENCE OTHER
arrive at different elevation angles with respect to the CALCULATION
I0
INTERF
(MES ETC.)
receiver and these signals experience different antenna C

gain levels. The combination of these effects and the C/(No+Io ) ANALYSIS
potential for interference is illustrated in Figure 1. COMPUTATION
AVAILABILITY
MODEL
OR
I0’ RESULTS

Receiver Receiver C/ N 0
Noise PSD Threshold
(N 0)

ANTENNA HIGH
Figure 2. CLIMAT Block Diagram
GAIN DIRECTION
ANTENNA LOW
GAIN DIRECTION
CARRIER-TO-NOISE DENSITY CALCULATION
INTERFERING SIGNALS INTENDED SIGNAL
(CROSS CORRELATION) (AUTO CORRELATION) From Figure 2, it can be seen that the satellite
availability model depends on two inputs. The first one
is based on the required carrier-to-noise density ratio
Carrier-to-noise density ratio (C/N0) can drop to
near or below acquisition threshold of 34 dB-Hz (C/N0), called receiver C/N0 threshold, for the receiver
to satisfactorily function. In the receiver code
Figure 1. Combined Effects of C/A Code Self- acquisition mode, a typical number used for the
Interference and Satellite-to-User Geometry threshold is 34 dB-Hz, and in the tracking mode a
typical value for the threshold is 30 dB-Hz. The second

2
American Institute of Aeronautics and Astronautics
 ' N
one is the available effective C / N 0 , which is equal to I0,CDMA = ∑ (α (P i i
CA
βCA )
C / ( N 0 + I 0 ) . For any satellite if the available
i=1

C / N 0' is less than 30 dB-Hz, that particular satellite is

I0,Intra = Equivalent Noise Density due to interference
not available for both code acquisition and tracking. On from all other code signals from all the visible
'
the other hand, if the available C / N 0 is less than 34 GPS Satellites. For receiving CA codes, other
dB-Hz, but greater than 30 dB-Hz, that particular code signals are from P, M and IM in GPS.
satellite is not available for code acquisition. So the key
∑ (α (P β + ( P ) β )
N
'
question is how do we calculate the available C / N 0 . I 0,Intra = i i
P
P i
M
M + PiIM β IM ) (3)
i=0

In this section the approach to the interference

calculation will be outlined. The basic approach used in
the analysis is to determine the effective carrier power Where
Downloaded by BEIHANG UNIVERSITY on April 3, 2020 | http://arc.aiaa.org | DOI: 10.2514/6.2004-3182

'
to one-sided noise spectral density ratio C / N 0 as a
' αi = Composite transmit/receive antenna gain, from
measure of the system performance. This C / N 0
ith satellite to the GPS user
measure is applicable to bit error rate (BER)
calculations and code acquisition, but not to code Px = Received power; x = CA, P, M or IM code.
tracking for which tracking error variance is a more
appropriate measure. Effective carrier-to-noise density N+1 = Number of visible satellites from a given user
'
ratio ( C / N 0 ) is shown in Equation (1) and can be location.
written as
β 's = Respective code despread factors (also know as
C C spectral separation coefficient) which is given
= (1) by
N'0 νN0+I0,CDMA+I0,Intra +I0,Inter +I0,External +I0,Margin

∫H
2
Where βX = BB (f) SX (f)SPN (f)df (4)
-∞

C = Code signal power from the desired GPS

Satellite which includes transmit/receive Where
antenna gains and receiver implementation
loss SX(f) is the two-sided unit-power spectral density (psd)
of the respective interfering signal from the same
N0 = Receiver thermal noise power spectral density system (e.g., GPS) and SPN(f) is the unit psd of the
signal being interfered with.
∞
ν = ∫ HBB(f) SPN(f)df ( ≤ 1)
2
(2) N

-∞
I 0,Inter = ∑ (α (P
i=1
i i
Inter
β Inter ) (5)
Where

HBB(f) is the baseband equivalent transfer function of Where

the transmitter and the receiver bandpass
filters and SPN(f) is the two-sided unit-power PiInter is the received power from the ith satellite in the
psd of the code being correlated at the other interfering system (e.g., Galileo) and βInter is given
receiver. by equation (5) with SInter(f) is the two-sided unit-power
spectral density (psd) of the respective interfering
I0,CDMA = Equivalent Noise Density due to interference signal from a different system (e.g., Galileo). Equations
from similar code signals from all other (5) and (6) can also be used to compute I0,External with
visible GPS Satellites. For example, for appropriate changes in the parameter values and I0,Margin
receiving CA code, is any margin mandated in the system operation.

3
American Institute of Aeronautics and Astronautics
 Table 1. Despread Factor β X in dB/Hz

SX ( f ) 1 MCPS
10 MCPS BOC(2,2) BOC(5,1) BOC(10,5) BOC(14,2) BOC(15,1)
S PN1 ( f ) NRZ
Random Random Random Random Random Random
Random

1 MCPS
NRZ –61.8* –70.2 –73.4 –82.4 –88.1 –91.2 –91.6
Random
10 MCPS
NRZ –70.2 –71.9 –71.1 –77.0 –80.9 –83.5 –84.2
Random
Downloaded by BEIHANG UNIVERSITY on April 3, 2020 | http://arc.aiaa.org | DOI: 10.2514/6.2004-3182

BOC(10,5)
–88.1 –80.9 –80.1 –87.9 –73.2 –82.6 –88.7
Random
* For C/A codes this value is much higher, as explained in the next section.
BOC – Binary Offset Carrier modulation used in modernized GPS military signal. Also, some versions of
this are proposed for future GPS and Galileo civilian use.

Spectral separation coefficients (SSC) or despread squarer, which are then used for calculating the I0,Intra
factors calculated using equation (4) for some sample term of the effective ( C / N 0' ) in Equation (1).
signals are given in Table 1. . For example, the
despread factor for interference from P-code (which is Received GPS
standard) to a C/A code signal is –70.2 dB/Hz and from Signal + Noise +
a C/A code assumed strictly random the despread factor Interference
I&D
is –61.8 dB-Hz. The interference noise density (I0) is X Filter ( )2
I2
simply obtained by adding the received interference C(t+T d)Cos(δ.) 2πfct
power to the corresponding despread factor. This + I&D Filter/

method works well for all the code combinations in + Threshold

+ Detector
Table 1 except for C/A codes for which the I0’s are Q2
much higher than can be obtained from Table 1 because X
I&D
Filter () 2

of the spectral line effects of the C/A code. Equations C(t+T d)Sin(.)2πfct
(3) and (4) can also be used to calculate the remaining
Local Oscillator/
interference psd in Equation (1) by choosing the Reference Code
appropriate psd for S X ( f ) . Generator

CODE ACQUISITION MODEL Figure 3. Noncoherent Code Acquisition Circuit

As mentioned earlier, code acquisition is the most

sensitive aspect of the system operation when it comes
to susceptibility to CDMA noise. This section
summarizes the CDMA noise calculation as applied to r(t) m(t) y(t) x(t) z(t)
RF/IF LPF •
2 LPF
the C/A code acquisition process, and provides sample (BIF) H1(f) H2(f)
numerical examples illustrating its susceptibility to a(t)
*

CDMA noise. A generic noncoherent code acquisition exp[+j2πfct]

*
Code-phase/Frequency Threshold
circuit is shown in Figure 3; the corresponding Select (Ta, fa) Device
baseband model used for analysis is shown in Figure 4. Figure 4. Baseband Code Acquisition Model
It is assumed that the search for the correct code phase
is performed serially with a step size of ½ code chip.
The analysis culminates in closed-form expressions for The two (baseband) input waveforms to the correlator
CDMA noise power spectral density at the output of the in Figure 4 can be expressed as

4
American Institute of Aeronautics and Astronautics
 r (t ) = s (t ) + n(t ) ignored in evaluating the CDMA noise. Since the
a (t ) = Ar ⋅ c0 (t − Ta ) ⋅ exp{ j ⋅ [2πf a t + ϕ a ]} bandwidth of the post-squarer low-pass filter H2(f) is
small—typically on the order of the navigation data
bandwidth—the peak of the PSD Sz(f) can be taken as
the “noise density level” at the input of the threshold
where a(t) denotes the locally generated reference code device in Figure 4. Therefore, the increase in this noise
waveform, s(t) denotes the received GPS signal density level (from the case of noise-alone to the case
(including M interference sources): of noise-plus-interference) provides a proportional
measure of the post-squarer CDMA noise. Taking the
square root of the post-squarer CDMA noise density
M value approximates the CDMA noise density level at
s (t ) = ∑ Ai ⋅ ci (t − Ti ) ⋅ exp{ j ⋅ [2π fi t + θ i (t − Ti ) + ϕ i ]} the correlator output (i.e., input of the squarer). This is a
i =0
good approximation when the signal-to-noise ratio is
very low, which is the case when the reference code
offset is more than a chip.
and n(t) denotes the zero-mean white Gaussian thermal
Downloaded by BEIHANG UNIVERSITY on April 3, 2020 | http://arc.aiaa.org | DOI: 10.2514/6.2004-3182

noise process with autocorrelation function

Numerical Examples
Rn (τ ) = 2 N 0 ⋅ BIF sinc( BIFτ ) ≈ 2 N 0 ⋅ δ (τ ).
Two sets of numerical examples are presented in this
The following assumptions are made in the analysis: section to illustrate the susceptibility of the code
acquisition process to CDMA noise. In both sets of
(A1) The propagation delays {Ti} and Doppler examples, the increase in post-squarer CDMA noise
shifts {fi} remain unchanged while dwelling in any density level is calculated using the equations
particular acquisition code-phase Ta. summarized in Table 2. All discrete PSD components at
f=0 are suppressed because they convey only DC
(A2) The pre-squarer low-pass filter h1(t) is real power.
and rejects all frequency components beyond one
half the C/A code repetition rate fR=1/TR. The first set of examples consists of three worst-case
scenarios described in Tables 3, 4, and 5, respectively.
(A3) The BPSK-modulated navigation data bits are These scenarios were taken at three different time
equally probable ±1, and the navigation data instants over a stretch of about 14 minutes during which
processes {θi(t)} are slow-varying with respect to the carrier-to-noise-density ratio C/N0 was degraded
the pre-squarer low-pass filter h1(t). significantly as observed from the output of the
CLIMAT. The descriptions of each scenario include the
(A4) The random carrier phases {ϕi} are total number of satellites visible to the receiver, as well
independent of the navigation data processes as the corresponding received power levels (including
{θi(t)}, and are independent and uniformly user antenna gain) in dBw, Doppler frequency shift in
hertz, and path delay in units of C/A code chip.
distributed as uniform over (−π, +π).
Assuming a thermal noise density level of N0=−201.5
(A5) The additive zero-mean white Gaussian dBw/Hz and a zero frequency offset between the
thermal noise process n(t) is independent of both desired and reference code waveforms, the resulting
increments in post-squarer CDMA noise density level
the navigation data processes {θi(t)} and the
as a function of the acquisition code-phase used in the
random carrier phases {ϕi}.
reference waveform are given in Figures 5, 6, and 7,
Table 2 summarizes the equations needed for
respectively. A common feature of these figures is that
calculating of the power spectral density (PSD) of the
the average CDMA noise increment is markedly lower
filtered squarer output z(t). As can be seen from these
than numerous “instantaneous” noise increments
equations, the PSD expression has three components:
irregularly scattered across the entire reference code-
an interference-on-interference (I×I) component, a
phase space. Also, as evidenced in Figures 8 and 9, the
noise-on-noise component (N×N), and an interference-
same feature prevails when the frequency offset
on-noise (I×N) component due to the interaction
between the desired code waveform and the reference
between the interfering code signals and the thermal
code waveform is non-zero.
noise. Furthermore, since the contributions from these
components vary, depending on the underlying Doppler
conditions, none of these components can be safely

5
American Institute of Aeronautics and Astronautics
 The second set of examples consists of a collection of received power of the desired code waveform is set at a
scenarios over a stretch of 2000 seconds involving three constant level that is 34 dB above a thermal noise
GPS satellites in view, one scenario per second. The density of N0=−200.5 dBw/Hz. The received power
Table 2. Summary of Equations to Calculate C/A Code CDMA Noise PSD
2
Sz ( f ) = Sx ( f ) ⋅ H2 ( f )
S x ( f ) = S x ,I×I ( f ) + S x ,I×N ( f ) + S x , N× N ( f )

∑ B ⋅ (S ( f − [ f ) ( ))
M −1 M
~ ~ ~ ~
S x ,I×I ( f ) = B 2 ⋅ δ ( f ) + 14 ⋅ ∑ 2
i ,n i ,n i ,a − f n ,a ] + S i ,n f + [ f i ,a − f n ,a ]
i = 0 n =i +1

( )
M
~ ~
S x ,I× N ( f ) = D 2 ⋅ δ ( f ) + ⋅ ∑ D ⋅ S b ( f − f i ,a ) + S b ( f + f i ,a )
1
2 i
2

i =0

S x , N× N ( f ) = 4 N 02 Ar4 ⋅ G 2 ⋅ δ ( f ) + 4 N 02 Ar4 ⋅ Λ( f )
Downloaded by BEIHANG UNIVERSITY on April 3, 2020 | http://arc.aiaa.org | DOI: 10.2514/6.2004-3182

S i ,n ( f ) = S i(,1n) ( f ) + S i(,2n) ( f ) + S i(,3n) ( f )

S b ( f ) = Tb ⋅ sinc ( fTb ) ∗ H1 ( f ) → 
2 b
2
b
( R
)
 1 − TTR ⋅ sinc 2 ( fTR ) + TTR ⋅ π ⋅ f1⋅T 2 ⋅ [1 − sinc(2 fTR )], ( ) f ≠0
 1 TR
1 − 3 ⋅ Tb , f =0


(
Λ ( f ) = H1 ( f ) ∗ H1 ( f ) →  R
2 2
R
) (
 T1 ⋅ π ⋅ f1⋅T 2 ⋅ [1 − sinc(2 fTR )], f ≠ 0 )
 3 ⋅ TR ,
2 1
f =0

2 M
 M
~ 2 ~ 2
B =  Ar2 ⋅ ∑ Ai2 ⋅ γ i ⋅ H1 ( f i ,a ) 
2
D 2 = 4 N 0 Ar4 ⋅ G ⋅ ∑ Ai2 ⋅ γ i H 1 ( f i ,a )
 i =0  i =0
~ 2
( 2 ~ 2
) ~ 2
Bi2,n = 2 Ar2 ⋅ Ai An ⋅ γ i ⋅ H1 ( f i ,a ) ⋅ γ n ⋅ H1 ( f n ,a ) Di2 = 4 N 0 Ar4 ⋅ Ai2 ⋅ γ i H 1 ( f i ,a )

G = ∫ h1 (u ) du = ∫ H 1 (v) dv →
2 2
( ) 1
TR

S i(,1n) ( f ) = 1
Tb (T~ ) ⋅ sinc ( f ⋅ T~ )
(min) 2
i ,n
2 (min)
i ,n

S i(,2n) ( f ) = 1
Tb (T~ − T~ ) ⋅ sinc ( f ⋅ (T~ − T~ ))
(max)
i ,n
(min) 2
i ,n
2 (max)
i ,n
(min)
i ,n

S i(,3n) ( f ) = 1
Tb (T − T~ ) ⋅ sinc ( f ⋅ (T − T~ ))
b
(max) 2
i ,n
2
b
(max)
i ,n
TR
γ i = T1 R ∫
0
ci (t − Ti )c0 (t − Ta ) ⋅ exp[+ j ⋅ 2π m i , a f R t ] dt

~
f i ,a = ( f i ,a − m
~ ⋅ f )∈ [ −
~
Ti = (Ti mod Tb )∈ [0, Tb )
fR fR
i ,a R 2 ,+ 2 )
~
Ti ,(min)
n
~ ~
= min Ti , Tn ( ) f i ,a = f i − f a
~
Ti ,(max)
n
~ ~
= max Ti , Tn ( ) ~ = int 1 ⋅ int
m i ,a 2
( ( ( )+ 1))
f i ,a
fR / 2

(Tb = Navigation Data Bit Duration) ( f R = 1 / TR = C/A Code Repetition Rate)

6
American Institute of Aeronautics and Astronautics
 from interfering satellites was set at 10 dB higher than
the power level of the desired signal. A Doppler shift
and a path delay are associated with each scenario, and
these Doppler/delay values are used to arrive at a figure
similar to those in Figures 5, 6, and 7, from which a
scenario-specific average and maximum increments in Table 4. Description of Scenario 5
post-squarer CDMA noise are obtained. Figure 10
Scenario #5 of "1+7"
collects the average/maximum CDMA noise increments
of all 2000 scenarios and displays them as a function of M := 7 i := 0 .. M

the scenario time line. Consistent with what was P := − 162.96 Delay := 80.62 ⋅ 1023 Doppler := 1875.13
0 0 0
observed in the first set of examples, the average
P := − 154.39 Delay := 68.98 ⋅ 1023 Doppler := − 1439.28
CDMA noise increments are markedly lower than the 1 1 1
maximum CDMA noise increments. Finally, as P := − 159.26
2
Delay
2
:= 75.83 ⋅ 1023 Doppler
2
:= − 3255.69
evidenced in Figure 11, the same observation holds
P := − 154.67 Delay := 69.93 ⋅ 1023 Doppler := 1870.60
Downloaded by BEIHANG UNIVERSITY on April 3, 2020 | http://arc.aiaa.org | DOI: 10.2514/6.2004-3182

even when the discrete PSD components are included in 3 3 3

the computation. P := − 157.58

4
Delay
4
:= 74.11 ⋅ 1023 Doppler
4
:= − 1202.71

P := − 163.25 Delay := 81.13 ⋅ 1023 Doppler := − 3103.15

Table 3. Description of Scenario 4 5 5 5

P := − 164.22 Delay := 83.14 ⋅ 1023 Doppler := − 3477.41

6 6 6
Scenario #4 of "1+7"
P := − 156.97 Delay := 73.48 ⋅ 1023 Doppler := 1869.01
7 7 7
M := 7 i := 0 .. M

P := − 163.31 Delay := 81.24 ⋅ 1023 Doppler := 2142.25

0 0 0

P := − 154.32 Delay := 65.59 ⋅ 1023 Doppler := − 1176.3

1 1 1

P := − 158.36 Delay := 74.90 ⋅ 1023 Doppler := − 3114.91

2 2 2

P := − 154.90 Delay := 70.54 ⋅ 1023 Doppler := 2122.13

3 3 3
k := 0 , 1 .. 2046
P := − 157.27 Delay := 73.79 ⋅ 1023 Doppler := − 880.66
4 4 4
12
P := − 162.69 Delay := 80.21 ⋅ 1023 Doppler := − 2915.6
5 5 5 11
Scenario #5 of Case ‘1+7’ Average = 1.3 dB
P := − 163.76 Delay := 82.10 ⋅ 1023 Doppler := − 3382.34 10
6 6 6 facq = f0
9
P := − 157.57 Delay := 74.09 ⋅ 1023 Doppler := 2139.57
7 7 7
Increased Noise Level (dB)

k := 0 , 1 .. 2046 2

1
12
0
11 0 256 512 768 1024 1280 1536 1792 2048
Scenario #4 of Case ‘1+7’ Average = 1.4 dB Code Phase Offset (Tc/2)
10
facq = f0
9

Figure 6. C/N0 Degradation at the Detector Input-

Increased Noise Level (dB)

7 Scenario 5
6

0
0 256 512 768 1024 1280 1536 1792 2048
Code Phase Offset (Tc/2)

Figure 5. C/N0 Degradation at the Detector Input-

Scenario 4

7
American Institute of Aeronautics and Astronautics
 Table 5. Description of Scenario 6 k := 0 , 1 .. 2046

12

11
Scenario #4 of Case ‘1+7’
10
facq = f0+1000 Hz
9

Increased Noise Level (dB)

8
Scenario #6 of "1+7" 7

M := 6 i := 0 .. M 6

P := −162.66 Delay := 80.70⋅ 1023 Doppler := 1590.90 4

0 0 0
3
P := −154.52 Delay := 69.46⋅ 1023 Doppler := −1694.27
1 1 1 2

P := −160.16 Delay := 76.89⋅ 1023 Doppler := −3381.86 1

2 2 2
0
0 256 512 768 1024 1280 1536 1792 2048
P := −154.48 Delay := 69.40⋅ 1023 Doppler := 1608.44 Code Phase Offset (Tc/2)
3 3 3

P := −157.99 Delay := 74.52⋅ 1023 Doppler := −1513.13

Figure 9. C/N0 Degradation at the Detector Input—
4 4 4 Scenario 4 with 1000 Hz Offset
Downloaded by BEIHANG UNIVERSITY on April 3, 2020 | http://arc.aiaa.org | DOI: 10.2514/6.2004-3182

P := −163.76 Delay := 82.10⋅ 1023 Doppler := −3267.80

5 5 5

P := −300.00 Delay := 0⋅ 1023 Doppler := 0 3.0

6 6 6

P := −156.1 Delay := 72.96⋅ 1023 Doppler := 1581.60 2.5

6 6 6 MAX

Increased Noise Density Level (dB)

2.0

k := 0 , 1 .. 2046
1.5
12

11
Scenario #6 of Case ‘1+7’ Average = 1.4 dB 1.0
10
facq = f0 AVG

9 0.5
Increased Noise Level (dB)

8
0.0
7
506000 506400 506800 507200 507600 508000
6 Tim e of Week (sec)

4 Figure 10. C/N0 Degradation at the Detector Input—

3 Scenario 34+10 dB, without DC
2

1 3.0

0
0 256 512 768 1024 1280 1536 1792 2048
Code Phase Offset (Tc/2) 2.5
Increased Noise Density Level (dB)

Figure 7. C/N0 Degradation at the Detector Input—

2.0
Scenario 6 MAX

1.5
k := 0 , 1 .. 2046

12 1.0

11
Scenario #4 of Case ‘1+7’ 0.5 AVG
10
facq = f0+100 Hz
9
0.0
Increased Noise Level (dB)

8 506000 506400 506800 507200 507600 508000

Tim e of Week (sec)
7

5
Figure 11. C/N0 Degradation at the Detector Input—
4
Scenario 34+10 dB, with DC
3

0
0 256 512 768 1024 1280 1536 1792 2048
Code Phase Offset (Tc/2)

Figure 8. C/N0 Degradation at the Detector Input—

Scenario 4 with 100 Hz Offset

8
American Institute of Aeronautics and Astronautics
 SystemView Simulation

A simulation of the block diagram of the baseband

noncoherent acquisition circuit shown earlier in Figure
3 is described in this section. The received signal,
corrupted by thermal noise and other interference
sources, is despread by the reference code with an
estimated code phase and carrier Doppler to form the
in-phase and quadrature components, which in turn are
low-pass filtered (Integrate & Dump Filter), squared
and added to form I2 + Q2 signal. This signal is further
filtered, if necessary, using another I & D filter and the
output is compared against a threshold in the detector to
decide whether the estimated code phase is within the
Figure 12. SystemView Simulation Block Diagram
Downloaded by BEIHANG UNIVERSITY on April 3, 2020 | http://arc.aiaa.org | DOI: 10.2514/6.2004-3182

tracking range of the code tracking loop. In this paper,

instead of measuring the detection and false-alarm rates
of the detector, effective carrier-to-noise density ratio
(C/N0’) is measured at the input of the detector from
which C/N0 at the correlator output (input of squarer) is
deduced. By comparing the C/N0 for noise-only case
with noise-plus-interference case, a C/N0 degradation
number is obtained.

A block diagram of a simulation of the above

acquisition circuit using SystemView software
(copyright of Elanix Corporation) is shown in Figure
12. Tokens 4, 12, 14, 15, 1, and 384 are used to
generate C/A code signal and Gaussian noise samples.
Tokens 7, 11, 193, and 383 are amplifiers to set the
gains for the desired C/N0 ratio. Token 427 is a DCMA
noise generator for which details are shown in Figure
13. For each visible satellite, received power levels
including user antenna gain in the direction of that
particular satellite, Doppler frequency shifts and path
delays as obtained from CLIMAT as explained before
are input to the simulation. Code acquisition circuit is
shown in Figure 14. Figure 13. SystemView Simulation—C/A Code CDMA
Noise Generators

Figure 14. SystemView Simulation Block Diagram—

Code Acquisition Circuit

9
American Institute of Aeronautics and Astronautics
 Simulation Results Conclusions

As an example, the desired signal and the interfering A two-step procedure is proposed to analyze the impact
signals were generated according to the scenario 6 of C/A code CDMA noise on the GPS C/A code
parameters as described in Table 6. Also, noise samples operation. Step 1 is to use the commonly used
corresponding to a noise spectral density of –201.5 '
methodology of calculating C / N 0 , ignoring the C/A
dBW/Hz were added to the signal. After correlation
with the reference code at some trial code phase offset, code spectral line effect. Step 2 is necessary only when
'
followed by I & D filtering and squaring, samples were the C / N 0 is different from C/No by more than 0.25
collected at the squarer output. The probability density dB or the C/No is very close to 34 dB-Hz. In Step 2, a
function (PDF) was calculated from these samples more comprehensive analysis model such as used in
using histograms and plotted in Figure 15. An CLIMAT should be utilized.
additional noise source was added at the input to
replace the CDMA noise from the interfering satellites, Aerospace has developed a comprehensive interference
and the noise density was set at a level to increase the analysis tool called CDMA Limited Interference Model
Downloaded by BEIHANG UNIVERSITY on April 3, 2020 | http://arc.aiaa.org | DOI: 10.2514/6.2004-3182

overall noise level by +1.25 dB and +2.5 dB over the and Analysis Tool (CLIMAT) to analyze the worst-case
quiescent noise level. The PDFs were calculated as impact of the C/A code CDMA noise on the code
described before and were plotted for comparison in acquisition and tracking performance of the GPS C/A
Figure 15. As can be seen from this figure, the PDFs codes. This tool combines geometric considerations of
corresponding to the C/A code CDMA noise (marked the user and the satellite orbits with the analytical
Sn+N+I(CA) in Figure 15) and a thermal noise increase modeling of the interference phenomenon. Analytical
of 1.25 dB (marked Sn+N(+1.25dB)) match very model validation has been done using a commercially
closely, indicating that an equivalent noise increase due available communication systems simulation software
to the C/A code CDMA noise in scenario 6 corresponds tool called SystemView (copyright by Elanix Inc.), and
to 1.25 dB. This number is more in line with the there is a good agreement between the analytical and
analytical results presented earlier using the closed- the simulation results.
form solution (see Figure 7).
SystemView
Numerical results obtained using the CLIMAT have
PDF of samples at the 50 Hz Filter Output
shown that under certain scenarios, the carrier-to-noise
120.e-3
20.e-3 120.e-3 220.e-3 320.e-3 420.e-3
density (C/N0) degradation due to worst-case CDMA
Sn+N Sn+N+I(CA) noise could exceed 1.5 dB, whereas traditional methods
100.e-3 used in the past indicate a degradation of less than 0.5
Sn+N(+1.25dB) dB. This 1 dB or more of the degradation difference
80.e-3
may not be very detrimental to the GPS operation in
Sn+N(+2.5dB) most of the applications but could be important enough
Bin Count

60.e-3

in safety-of-life type applications. Also, when GPS

40.e-3
frequency band sharing with dissimilar services is
considered over 1 dB of C/N0 degradation may make all
20.e-3
the difference between satisfying and not satisfying the
mission requirement.
0

20.e-3 120.e-3 220.e-3

Amplitude
320.e-3 420.e-3 Although it is desirable to use CLIMAT in all sharing
studies, because of the complexity of the model
Figure 15. PDF of Samples at Threshold Device involved, it is suggested that a tiered approach be taken
Input—SystemView Simulation in the interference analysis. In the first step, C/N0
degradation can be calculated using the more traditional
but simple method (for example, convolving GPS and
interfering signal power spectral densities to determine
the peak interference contribution). If the degradation is
less than 0.5 dB and the resulting C/N0 is above 36 dB-
Hz, there is no need to run CLIMAT to determine the
degradation with any further precision. If, however, the
degradation is more than 0.5 dB or the resulting C/N0 is
less than 36 dB-Hz, a more comprehensive study
should be done.

10
American Institute of Aeronautics and Astronautics
 CLIMAT was developed primarily to study C/A code-
related issues, but it can very easily be modified to
analyze other codes such as being proposed for L2C,
L5, or any other codes.

References

[1] S. Lazar, S. Raghavan, and D. Turner, “GPS

Spectrum: Sharing or Encroachment,” GPS World
11, no. 9 (September 2000): 46–56.

[2] S. Raghavan et al., “The CDMA Limit of C/A

Codes in GPS Applications-Analysis and
Laboratory Test Results,” in Proceedings of the
ION GPS-99, 12th International Technical
Downloaded by BEIHANG UNIVERSITY on April 3, 2020 | http://arc.aiaa.org | DOI: 10.2514/6.2004-3182

Meeting of the Satellite Division of the Institute

of Navigation, Nashville, TN, September 15,
1999.

[3] Bradford W. Parkinson and James J. Spilker Jr.,

eds., Global Positioning System: Theory and
Applications, Volume I (Volume 163 in Progress
in Astronautics and Aeronautics) (Washington,
DC: AIAA, 1996), page 63, Volume 1.

[4] R. L. Peterson, R. E. Ziemer, and D. E. Borth,

Introduction to Spread Spectrum Communications
(Englewood Cliffs, NJ: Prentice Hall, 1995).

[5] Kuang Tsai, “Cross-Code Interference for GPS

C/A Code Acquisition,” Aerospace IOC, February
26, 2002.

[6] A. J. Van Dierendonck, “Scenario 34+10 dB”

description, personal communication.

11
American Institute of Aeronautics and Astronautics