Optimisation and sensitivity analysis of GPS

receiver tracking loops in dynamic environments

D.-J.Jwo

Abstract: For a GPS receiver, decreasing the receiver tracking loop bandwidth reduces the
probability of loss of lock if there are no vehicle dynamics. However, reduced bandwidth increases
tracking errors due to dynamics. Beyond a certain limit it causes a serious degradation in the
dynamic tracking performance. Therefore, there is involvement of a tradeoff between two
opposing considerations: narrow tracking loop bandwidths are desired for filtering noise due to
thermal effects, but wide tracking loop bandwidths are desired to permit tracking of vehicle
dynamics. Optimal tracking loop bandwidths, which yield the minimum errors in a certain
dynamics environment, are first investigated. The linear Kalman filter is employed as the optimal
estimator. The covariance for the arbitrary gain model is solved and applied to the sensitivity
analysis for investigating error growth due to incorrect noise level estimate. Theoretical results are
verified by numerical simulation, and results from both approaches are in very good agreement.

1 Introduction dynamics. Based on knowledge of the errors, the theore-

The tracking errors of a receivcr operating on the GPS code
and carrier include two major components: noise error,
caused by thermal noise; and transient error, caused by
imperfectly tracking the vehicle dynamics. Selection of the
baseband processor design for a GPS receiver always
involves a tradeoff between two opposing considerations.
Narrow tracking-loop bandwidths are desircd for filtering
noise due to thermal effects or jamming, however, wide
tracking-loop bandwidths are desired to permit tracking
signal Doppler shifts induced by vehicle/user dynamics.
Either the carrier loop or the code loop is usually
designed to select a bandwidth which produces tracking
errors under maximum dynamics approximately equal to
the lock limit of the loop. When the GPS signal power is
limited, the tendency would seem to make the receiver
tracking-loop bandwidth narrower. However, this increases
the probability that the tracking loop will lose lock owing
to vehicle/user dynamics. Thus, there is a fundamental
system limitation of tracking-loop threshold when consid-
ering both low carrier-to-noise ratio and user dynamics at
thc same time. It is therefore important to analyse the error
characteristics for determining the optimal loop bandwidth
that minimises the total tracking error.
Receiver noise models for predicting the thermal noise
,jitters have been presented, for example [l-71. Dynamics
stress errors can also be accurately predicted [4, 6, 71. The
summation of these two major error components has a
minimum value for certain carrier-to-noise ratio and user
IEE. 2001
IEE Pmcerdings online no. 20010429
DUI: 10.1049/ip-rsn:20010429
Paper first received 23rd October 2000 and in revised form 27th March
2001
The author is with the Department of Guidance and Communications
Technology, National Taiwan Ocean University, 2 Pei Ning Rd., Keelung
202-04, Taiwan, Republic of China
IEE Proc.-Rudu,: Sonar Nuvig,. Vol. 148, No. 4, August 2001
tical prediction of optimal bandwidth for minimum track-
ing error can be determined. The environment concerned is
the case of limited GPS signal power or jamming for the
dynamics user without other information aiding. How the
incorrect parameters (i.e. departure from the design point:
could be intentional or unintentional) influence the error
growth, referred to as the ‘sensitivity analysis’, is consid-
ered. The sensitivity analysis involves the incorrect esti-
mate of received signal carrier-to-noise ratio. A theoretical
approach and numerical simulation are performed for
verification.
2 GPS receiver tracking loops
The GPS receiver contains a code tracking loop and a
carrier tracking loop for tracking the Doppler-shifted
carrier. The pseudorange obtained from the code tracking
loop provides a position fix; the pseudorange rate estimate
obtained from the carrier tracking loop provides a velocity
fix. The receiver carrier tracking loop is more sensitive to
dynamics due to the fact that it tracks a much higher
frequency signal than a code tracking loop. If the carrier
tracking loop loses lock during a dynamic manoeuvre, the
code tracking loop will usually lose lock subsequently.
Some designs use the carrier loop to track the dynamics
and provide the code loop with a prior knowledge of the
dynamics such that the code loop will not see the full
dynamics. The external navigation source, such as inertial
velocity, can also be utilised to aid the tracking loops for
removing most of the dynamics stress error such that a
smaller bandwidth could be used. More information
regarding the inertial velocity aiding can be found in
[9, lo]. The architecture of the simplified GPS receiver
tracking loops is shown in Fig. I .
2.1 Transfer functions of tracking loops
A generalised block diagram of a tracking loop that is
applicable for analysis of both carrier and code loops is
241
 delay discrim.
.............................
code VCO change of acceleration) input; fourth-order loops are insen-
sitive to jerk. It is rare that a loop is constructed with an
order higher than third. More information on tracking loop
related topics is provided in [7].

2.2 Equivalent noise bandwidth

(taid) (inertial velocity aiding) The single-side equivalent noise bandwidth, in Hz, for a
tracking loop with transfer function H(jrc,) is expressed as
phase det.
...............................
(3)

carrier VCO where ~ 0 = 2 n f and the magnitude of the frequency

response is

I H(J'LII)
1' = [ ~ ( j w ) ~ ( - j o ~ ) l (4)
Fig. 1 Rloclc diagram ?t'.yimpl$ed GPS receiver tmclcing loop,^ 191
Analytical solution for the tracking loop bandwidths, by
implementing eqn. 3, might be tedious. R. S. Philips' table
shown in Fig. 2. Thc closed-loop transfer function of the of contour integrals can be employed for simpler computa-
tracking loop is tion

where the open-loop transfer function is

where
c(s) = C,,-p"-' + C,i-Z"n-2 + + CO ' ' '

KO represents the voltage-controlled oscillator (VCO) gain a(s) = a,,.s" + L 2 , , - , S n - l + . .

' +U(,
factor, K,, is the phase-detector/delay discriminator gain
Fig. 3 provides solutions for such typc of intcgration up to
factor. The transfer function F(s) reprcsents the loop filter
fourth order. Solutions for higher orders can be found in
of a tracking loop. The loop filter is usually a low-pass
[8]. Fig. 4 provides the loop filters, closed-loop transfer
filter used to suppress the noise and high-frequency signal
functions, and equivalent noise bandwidths, for different
components from the phase-detector/delay discriminator
orders of receiver tracking loops.
and provide a DC-controlled signal for the VCO. A first-
order tracking loop is obtained when F(.Y)= 1 There are
usually three options for selecting the loop filters in a 3 Receiver error behaviour
second-order loop: a (simple) lag filter, an activc filter or a
passive filter. A second-order loop with lag filter is usually The tracking errors of the receiver operating on thc GPS
referred to as a modified first-order loop rather than a code and carrier have two major components: noise error
genuinc second-order loop. The second-order tracking caused by thermal noise, and transient error caused by
loop with passivc filter when 1/ K << z2 will be nearly the imperfectly tracking the user dynamics.
samc as the one with active filter. Third orders are insensi-
tive to acceleration, and optimal for constant jerk (rate of
3.7 Thermal noise
Closed-form expressions for the tracking error due to
G(s) thermal noise are derived undcr the assumption of infinite
.
...
...
...
...
...
...
...
signal bandwidth [3]. The onc-sigma errors of thermal
noises with regard to diffcrent types of code tracking loops
are as follows:
Coherent
a

Earlynzinus-late power

Dot-product

r B,d
b
Fig. 2 iMulhematicul modeh of'linenri.wd tracking loop where otDLL is the standard deviation of tracking error in
a In terms of original circuit parameters units of PRN chips, B, is thc single-sidcd code tracking-
0 In terms of natural frequency loop bandwidth in Hz, T is thc predetection integration
242 IEE Pmc.-Radar: Soiior h'avig., h l , 145, No. 4,Airgzist 2001
 u(s) = a,s" + a,,- + + a"
,.Fl ' ' '
c:,
I , =-
2a"(l ,
time (PIT) in scconds, U' is the early-to-late correlator
spacing normalised with respect to one chip
U' = I for time-shared tau-dithcrcd early-late correlator
= f for dedicated early and late correlator
and clno is the carricr-to-noise ratio value
clno = (SNR)(BL)(ratio-Hz)
CllY, = 10 log,o(c/no) (dB-Hz)
The second term in parentheses is known as the 'squaring
loss', Most GPS receivers employ a noncohercnt delay-
lock loop (DLL) for code tracking and a Costas-type
phase-lock loop (PLL) for tracking the Doppler-shifted
carrier. The thermal noise jitter for the commonly used
tau-dithered early-late DLL with d = I , is
The P-code and CIA-code, with the chipping rate (chips
per sec) of 10.23 and 1.023 Mbit/s, correspond to a chip
width i,, of 29.305 and 293.05 m, respectively.
Thermal noise for the PLL, when implemented in thc
form of a Costas-type loop, is approximated by
- [x(+-
c/no
1
2Tc/no
(degrees)
The GPS has a 50 Hz navigation data message bit rate; the
predetection integration timc i s usually the period of a
navigation data bit, 20 ins. The wavclcngths A, for L , and
L, carricrs, due to the link frequencies 1575.42 and
1227.6 MHz, are 0.193 and 0.2442 m, respectively.
IEE Pmc.-Rudur: Sonar Nuvavig.. Vol. 148, ,!Vo 4,Azrg~ist2001
For the DLL the thermal noise is indepcndent of
tracking-loop order; for the PLL, the thermal noise jitter
is not directly dependent on the loop order, too [4]. The
C/NOs of the GPS with good signal power typically range
from 35-55 dB-Hz, so tracking errors generally run on the
lower end of the range. For the nominal range of C/No
(35 dB-Hz and above) the squaring loss is usually consid-
ered negligible. Typically, the unaided code tracking-loop
bandwidths are in the range of 1-4 Hz and unaided carrier
tracking-loop bandwidths are in the range of 5-1 5 Hz. Fig.
5 shows one-sigma thermal jitters of the early-minus-late
noncoherent DLL and the Costas-type PLL, respectively.
The maxiinum bandwidths allowed for maintaining lock in
various GIN, environments arc shown in Fig. 6.
3.2 Dynamic stress errors
A small steady-state crror is usually desired and is consid-
ered as the criterion of good tracking performance. The
transfer function representing the tracking crror is
The steady-state error can be evaluated by means of the
final value theorem of the Laplace transforms
lirn e(t) = liin sE(s) = lim s[ 1 - H(s)] (10)
t ix s+o .v+ 0
and thus
(Xb)
The loop order is sensitive to the same order of dynamics,
e.g. first order to velocity stress, second order to accelera-
tion stress, and third order to jerk stress. The first-order
loop is suitable for a user position that varies in a random
walk manner (white noise velocity); the second-order loop
is suitable for a user velocity that varies in a random walk
manncr (white noise acceleration); the third-order loop is
suitable for a user acceleration that varies in a random walk
manner (white noise jerk). For example, the dynamic
condition resulting in a second-order loop tracking error
i s constant acceleration. There is no steady-state transient
243
Loop order 1st order 2nd order (active lag-lead) 3rd order

Formulation 1: tracking loop description in terms of original circuit parameters

F(s) 1 z2s + 1
7,s
(22)'

~
K
s+K

BL K
-
4

Formulation 11: tracking loop description in tcrms of natural frcquency

F(s) -
0, awns + Wf bqp2 + aw;s + w:
K Ks Ks2

Q),

5 + w,
BL a2 +I ah2 + a2 - b
4 4a 4(ab - 1)

Relationship between above two formulations

ID, = K(= K,,K,j)
UW, zz
72
--K
71
bo, = () 2
K

Fig. 4 Loop.filtrr.s, closed-loop tmn.Tfir,functions, and equivulent noise bandwidths E , (11~)for different ordim uf receiver tracking loops

error for a second-order loop with constant velocity input
and is optimal for constant acceleration input. The limita-
tion exists during the transient period when going from no
acceleration to constant acceleration.
From eqn. 9 the error transfer functiols due to user
dynamics are
3.2.7 Code tracking: Peak errors caused by a range
rate, range acceleration, and jerk step in the third-order
loop can be approximated by the steady-state response of a
first-, second-, and third-order loop caused by a range rate,
rangc acceleration, and jerk step function, respectively. The
three-sigma steady-state errors can be written as (in units
of chips)

(first order) (12a)

d2z/dt2 7(')
-
-
CO;
= w; (second order) (13h)

(second order) (12b) -

d3z/dt3
33 s3 - = CO: (third order) (13c)
-
-
Ly3 + bWns2 + a'o:s (= s3 + K1s2+ Kzs + K3 ) where d"z/df', = 1 . . , 3 , is the maximum linc-of-sight
(third order) (12c) dynamics. For example, d3r/dt3 represents the maximum
line-of-sight jerk dynamics. For a dynamic user platform,
the closed-loop bandwidth must be sufficiently wide so as
and the dynamics stress errors are determined as follows. to track the transient variation of delay against time that is
244 IEE Proc.-Radar, Sunar Navig.. Vol. 14& No, 4.AuguN 2001
 have lost lock. The peak errors caused by a phase rate
(frequency), phase acceleration (rrequency rate), and phase
jerk (frequency acccleration) step in the loop can be
approximated by the steady-statc rcsponse of a first-,
second-, or third-order loop, respectively, and the steady-
state errors are

=A f
_ . (second order)

4
=-
47 (third order) ( 14c)
m,l
a
where Afand c$~, are in units o f Hz and cycles. From the
20 -
rule-of-thumb 14, 51, the three-sigma code noise should not
exceed 1/2 chip of thc chip length and the carrier noise
should not cxcccd 45' of the wavelength for maintaining
16-
... ............
. \\ threshold signal lock. Therefore the onc-sigma code noise should not
exceed 1/6 chip of thc chip length and the carrier noise
m
U
should not exceed 15" of thc wavelength, i.e.
12 -

0- and
r

4-
opLL= d G nf + o: + dJ3 -< I5 (degrees) (1 6)

for DLL and PLL, respectively. The parameters CT~,and nA

0 I are the one-sigma vibration induced oscillator jitter and
Allan variance-induced oscillator j itter, respectively, which
CiNO,dB.Hz
b
arc small when compared with the other two and are
negligible without loss of generality.
Fig. 5 Onesigma t/zern?al,ji//er,~
a Early-minus-late noncoherent delay lock loop
b Costas-lypc phasc-lock loop 4 Optimal and subtoptimal linear solutions

When designing a system, there may be many possible

250 r results depending on the criteria of performance, the nature

Fig. 6 Maximum hundwidths allowed fiv mriintuining lock in vririuirs
carrier-lo-nni,~ratio C / N o envirunments.fix DLL imd P1.L
caused by user-satellite range dynamics without creating
large transient errors. In the absence of any kind o f aiding,
receivers that experience significant accelerations, such as
aircraft, must employ wider bandwidths and/or higher-
order tracking loops than stationary receivers.
3.2.2 Carrier tracking: If the error should become too
large that the VCO skips cycles, the loop is considered to
I M Proc -Rudiii; Srmar Nrivig, Pill. 14X, ~VIJ,4, , 4 1 i ~ i r . v2001
/ 245
of the input signal, and restrictions placed on loop config-
uration [7]. There is usually no unique optimal result that
applies undcr all conditions. Improvement in one area of
performance is usually at ttic expense of degrading the
other and therefore some compromise between the two is
always necessary.
The term 'optimal' here is meant the one with minimum
error variance. Linear approximation is made when imple-
menting thc analysis to the tracking loop. In addition, the
noise is assumed white. The propagation of the error for a
Kalman filter can be described by the Riccati equation
P = FP + P F +~ GQG" -P H ~ R - ~ H P (17)
When the system reaches steady state, P = 0, the equation
becomes an algebraic Riccati equation (ARE), which can
be solved for the steady-state minimum covariance matrix
and then the optimal Kalman gain matrix
K , =P,H~R-' (18)
The steady-state Kalinan filter uses gains derivcd from the
steady-state covariance and provides suboptimal solutions.
The example used for illustration is for an airplane that
rolls into a turn or a step input is applied to thc elevator
controls, adopted from [O]. Such a manoeuvre may be
modelled as a ramp in acceleration, starting at zero accel-
eration at time to and lcvclling off at constant value
acceleration A at time t , . This profile is describcd in
terms of the acceleration derivative as a pulsc in jerk,
ti - to wide and with amplitude A / ( t , - foj. The accelera-
tion profile for a high performance aircraft subjected to a
step stick elevator input at Mach 0.8 is very close to a ramp
levelling off after 0.6 s at 6 G acceleration. This manocuvre
is typical for a high performance jet aircraft operating in a
terrain-following mode. For such a case, thc parameter
jerk = 10 G/s = 98.1 m/s3 i s used as the dynamic input to
the tracking loop.
4.2 Suboptimal solutions
Referring to Fig. 3, the transfer functions in terms of
original circuit parameters can be represented in tcrins of
natural frcquency, which provides better insight on the
relationship between 5, and q1. The suboptimal solutions
and related properties are described as follows.
4.2.7 First order: Both thc gcncral and optimal forms
of the transfer function and bandwidth have the same result

4.1 Typical optimal gain and variance profiles

The time-varying crror covariance and Kalman gain matrix
can bc obtained using the fourth-order Runge-Kutta inte- The minimum variance and Kalman gain, respcctively, are
grator. To observe the typical gain and variance profiles,
assume that the signal powcr condition is at C/No= 30 dB- P , = ( Q I ? ) " ~ ;K , = K , = (Q/R)"* = a),,,
Hz, the dynamic equation is modelled as a 10 G/s jerk
LO,, = (Q/R)1'2
motion corrupted by white noise i.e. x = 98. I M) where +
w-N(0, Q ) and Q = 10, and the initial covariance is
4.2.2 Second order: The general form of transfer

Po=
1000
0
0
0
1000
0
:]
1000
function and handwidth for thc second-order loop are

Based on these assumptions the optimal bandwidth respectively. The minimum covariance and Kalinan gain
B, = 12.2258 Hz, and consequently, thc minimum matrices, respectively, are
measurement noise R = 105.2581nz2 are obtained
(explained in Section 5 ) . Fig. 7 shows the typical profiles
of Kalman gains and error variances for a third-order GPS
receiver tracking loops.

r
8-

The performancc is optimal when

6 -

resulting in the bandwidth B, = 3wn/42/2.

4.2.3 Third order: The general forms of the transfer

function and bandwidth for the thc third-order loop are

H(.r) =
+
K , s 2 +K2s Ki
-
bf,,,.? awts
~ + + (U; ,

a .Y3 + K ,.I2 + K,S + K; J.3 + /m,,s' + am,+ + OJ: '

BL =
+
ab2 a2 - h
w,
1200 4(ab - 1)
respectively. The minimum covariancc and Kalinan gain
1 000

80C
N
E
ai
e
._m
60C

i3
40C

20c

time, s
b

Fig. 7 Qpicul profile.?,for tlzidoTder tmrking IOO/J,S

a Kalman gains
b Error vai-kdii-lance resulting in the bandwidth BL 5tu1,/6.
246 IEE Proc.-Rurlui; Suizur Nuvig., W . 145, No. 4, AugLl.sf Z(J01
 Based on the preceding discussion, the values of B,/cir, or
are

-BL -
_ 1
- (first order)
Wn 4
--
-
a' +I (second order)

-
4a
+
ab2 a2 - b
(third order)
=0.3024 , ) (first order) (240)
4(ah - 1) ~

2c/no I+--
Tcln, :A
If defining the parameter CI = B,/cri,,, the dynamic stress
errors with minimum mean square error can then be
obtained utilising a and eqn. 14

A!
= - = 0.25-
A? (first order) (1 9a)
*<,
c01, BL =""I54 ( 2,
' (second order) (24b)
'fx 0.2812-A !
~

- (2); (second order) 2c/n, i

;
(0; 42/2 R B'L

- $ (i)3$= x 0.5787-Af
52
(third order) (19c)

= 1.0424 ' ) (third order) (24c)

2c/no
~ I+- Tc/no 1;
5 Optimisation of loop bandwidth

Optimal bandwidth is a function of the input carrier-to- 180r

noise ratio. The variable-gain loop is capable of performing
an explicit estimate of C/No as a criterion for adjusting the
loop bandwidths to determine an optimal bandwidth that
minimises the total loop tracking error
5- 120
e
100
c

-
P
U
2
'? 60
80

wherc dn)= d n z / d Y . To minimise the total error the loop is 7 threshold

designcd to bc capable of measuring C/No and adjusting 40
its bandwidth for optimal performance. It is seen that total
DLL tracking error is a function of several parameters: 20

oDLL =,f(B,, C/Nn, 1; d")),and B, value has opposite 0

0 2 4 6 8 10
influence on o , ) ~ ,The
~ . optimal bandwidth is determined
bandwidth, Hz
by differentiating mDLL with respect to BL and setting the a
result equal to zero
3'0 i
which yields

The optimal loop bandwidth is found to be

0.5 I I
20 25 30 35
C/N,, dE.Hz
b

Fig. 8 Total tracking error performance and optimal bandwidth f o r

minimum errors and bandwidth lower hound f i r maintaining luck in
various carrier-to-noise ratio envir-onmentvjbr third-order DLL (CIA rode)

IRE Pror.-Radal; Sonur Nuvig., Vol. 148, No 4, August 2001 247

 0.020 r I 6 Sensitivity analysis

Choice of optimal bandwidth strongly depends on modula-

I
E
L-
0.0161

? 0.01 ~
c I\ / Top
tion parameters; it is a priori knowledge of message
statistics that permits threshold reduction. The adaptive
loop performs selection of the desired bandwidth from a
m table of precomputed values based on the noise level C,",j
9 estimate, and control logic to modify the bandwidths and
2 0.00; prevent transients. As shown before, received C/No is the
key parameter in system performance analysis. In some
25 situation, there is an incorrect estimate of noise level; in
k' some applications, there is no need to adjust the loop to
o'ool0
30
35 40, 45.
\ \
I
attain exactly the best performance. The sensitivity analy-
sis usually involves in how the incorrect parameters
0 2 4 6 8 10 (departure from the design point) selected influence on
bandwidth, Hz the error growth.
a
The filter gain in the error covariance eqn. 17 does not
appear explicitly. Another form of error covariance propa-
gation is described now. The error covariance relationships
for a discrete filter with the same structure as the Kalman
filter, but with an arbitrary gain matrix are written as

2'ol
1.5

1 .o
LB
20 25 30
CIN,, dB.Hz
b
Fig. 9 Total tracking err07 peTformance and optimal handwidth for
minimum errors and handwidth lower hound for maintaining lock in
various carrier-to-noise ratio environments.for third-order PLL (L, carrier)
Using the same procedure for deriving the optimal PLL
loop bandwidth, we have
and
Figs. 8 and 9 provide examples on the tracking perfor-
mance for the third-order DLL and PLL, respectively.
Since the carrier-tracking accuracy is excellent even at
wide bandwidths, its velocity measurements can be used
to aid the code-tracking loop. Using carrier-aided code
techniques makes the dynamic stress error for the code-
tracking loop a small-order effect that usually can be
assumed negligible. If inertial information is delivered to
the tracking loops, the required minimum bandwidth can
be reduced without the penalty of increasing dynamics
errors [6, 9 , 101.
248
,
which can be described in a single differential equation for
35
the continuous filter
P = (F - KH)P + P(F -K + +
H ) ~ G Q G ~ K R K ~ (29)
This equation defines the error covariance for the filter with
a general filter gain matrix K , which can be solved for the
covariance of an arbitrary gain model. The sensitivity
analysis can be conveniently implemented by using this
representation. Taking the partial derivative of P, with
respect to K, using aPm/aK=0 for a minimum gives the
same result as eqn. 17.
If the fixed-gain matrix K of a filter has been designed
for particular values of Q and R , the steady-state error
covariance will vary linearly with the actual process noise
spectral density or measurement error spectral density. If
the actual noise variances are assumed fixed and the design
values of Q and R are varied, quite different curves result.
Any deviation of the design variances, and consequently K ,
from the correct values will cause an increase in the filter
error variance (a consequence of the optimality of the
filter). More information on sensitivity analysis is
contained in [ 111.
The error covariance for the filter with arbitrary gain
model, can be obtained by using eqn. 29
First order
P = -2KPfK2R+Q
Second order
I&? Proc.-Radul; Sonar Navig., Vol. 148, No. 4, August 2001
 incorrect estimate of the CIN, (i.e. design CIN,, deviating
from the actual ratio). The optimal bandwidth
(BL)DLLapfimum is first determined by a given estimate of
CjN,. With knowledge of CjN, and the related optimal
bandwidth, the measuremcnt noise variance R , and conse-
quently the Kalman gain matrix K , are determined. An
incorrect estimate of GIN, gives an incorrect gain matrix
and results in error growth.
The information flow chart for the sensitivity analysis is
shown in Fig. 10. Procedures for theoretical approach are
By setting P = 0, the solutions are found to be as follows. After input of design and actual C,”, and
dynamics information, the optimal bandwidths for design
P, =-
Q + K*R (first order) (33) and actual environments, i.e. (B,),,,, and (B,),,,, , are
2K subsequently determined by eqns. 24 (for DLL) and 26 (for
PLL). Once the optimal bandwidths are determined, noise
information, i.e. Rdcsignand R,,,, can be estimated from
KtQ + K2Q + K:R eqns. 7 (for DLL) and 8 (for PLL). Consequently the
optimal gain matrices, i.e. KdZsjgll
and K,,,, are determined
(second order) (34) by eqns. 17 and 18. Finally, the error covariance is
determined by eqns. 33 and 35. Error variance obtained
by actual measurement noise R,,, with K,,, (based on
(third order) ( 3 5 ) correct estimate on C,”,) has a minimum value while error
variance obtained by actual measurement noise R,.,,, with

where
P I ] , = K:K2K3R - KlKiR +K:KjR +KIQ
P12, = KI(K.K,R
P13m
+ KI Q>
( K j R + Q)(K1K2 - K3)
350L
300

+
P , ~ , = K;K,R K:Q K:R K ~ Q + +
p23 oo = K2 (K2KiR + K? Q)
P33, = K & R + K ~ K ~ Q + K I K ~ Q K&Q
- 5 I \ \ -

A = ZK3(KiK2 - K3) 5 150

The sensitivity curve of the error growth is predicted using 100

eqns. 33-35 for first- to third-order loops, respectively. The
terms ‘sensitivity’ here involves the error growth due to
I
20 24 28 32 36 40
Design
__ -
Actual
C/NO,dB.Hz
design C/NO actual C/No Fig. 11 Sensitioity anall;sisfor third-order DLL (CIA code)
& dynamics information & dynamics information
~ thcorctical
+ simulation, C/N;;= 28 dB-Hz
0 simulation, 30 dB-Hz
A simulation. 32 dB-Hz
calculated by actual optimal
bandwidth, (BL)design eqns. 24 for DLL bandwidth, (BL)true
& 26 for PLL
4.0

3.5
design meas. noise, calculated by
eqns. 7 for DLL
actual meas. noise, -
& 8 for PLL
hue %
;; 3.0
I

N
E

I 12 2.5 ai
.-
design gain matrix Kalman gain matrix J
&rue 2.0
s!
z
1.5
I I

covariance matrix covariance matrix 1.0

20 24 28 32 36 40
obtained by theoretical obtained by theoretical UNo, dB.Hz
approach and numerical approach and numerical
simulation simulation Fig. 12 Sen.sitivi@ unalysisfor third-order PLL (L, carrier)
- theoretical
+ simulation, C/.q = 28 &-Hz
Fig. 10 Informarion flnw chart for implementing sensitivity analysis 0 simulation, 30 dB-Hz
when incorrect carrier-to-noise ratio information is utilised A simulation, 32 dB-Hz

IEE Prac.-Radar. Sonar Navig., Yol. 148. No. 4, August 2001 249
Kdesrgn(duc to originally incorrect estimate on C/N,)
results in error growth. The numerical siinulation essen-
tially follows the same procedures as theoretical approach
except that random numbers generated by Matlab are
employed. The error variances are determined by analysing
the output data from the filter. Figs. 11 and 12 give
example on the sensitivity for the third-order DLL (C/A
code) and PLL ( L , carrier), respectively. The theoretical
results, represented by solid lines, and the simulation
results, represented by symbols, are in very good agree-
ment.
7 Conclusions
Analysis of optimisation and sensitivity of GPS receiver
tracking loops has been presented. The maximum band-
width threshold is essentially governed by the thermal
noise jitter; however, the minimum bandwidth threshold
is mainly govemed by the dynamic stress error. Tracking
loops are considered without internal or external aiding
such that the influence of the dynamics stress error
component on the total tracking error is significant. A
theoretical approach for predicting optimal loop bandwidth
has been derived. Sensitivity analysis of the error growth
due to an incorrect estimate of the carrier-to-noise ratio
was conducted. Mathematical derivation of theoretical
results of the tracking loop performance, including opti-
misation and sensitivity, has been performed from first to
third orders. Numerical examples on third-order loops for
high dynamics applications have been provided. Both
250
theoretical and simulation results are in very good agree-
ment. Implementation on the other orders of loops can be
done without difficulty based on the same procedure.
8 References
1 SPILKER, J.J.: ‘Fundamental of signal tracking theory’ in PARKTN-
SON, B.W, SPILKER, J.J., PENINA, A., andENGA, P. (Eds.): ‘Global
positioning system: theory and applications’, vol. 1 (AIAA, Washington
UC, 1996), Chap. 7
2 VAN DIEREDONCK, A.J.: ‘GPS receiver’ in PARKINSON, B.W.,
SPILKER, J.J., PENINA, A., and ENGA, P. (Eds.): ‘Global positioning
system: theory and applications’, vol. 1 (AIAA, Washington DC, 1996),
Chap. 8
3 BRAASCH, M.S., and VAN DIEREDONCK, A.J.: ‘GPS receiver
architecture and measurements’, Pmc. IEEE, 1997, 87, (I), pp. 48-64
4 WARD, PW: ‘Satcllite signal acquisition and tracking’ in KAPLAN,
E D . (Ed.): ‘Understanding GPS: - principlcs
. and anplications’
.. (Artech
House, MA, 19Y6), Chap.-5
5 WARD, P.W.: ‘GPS receivcr RF interfcrence monitoring, mitigation, and
analysis techniques’, Nuvig. J: Ins(. Nuvig., 1994-1995,41, (4), pp. 367-
791
6
~ , I
CAHN, C.R.: ‘Spread spectrum applications and state-of-thc-art equip-
ments‘ in ‘Spread spcctnnn communications’, AGARD lecturc serics 58,
1973
7 GARDNER, F.M.: ‘Phaselock techniques’ (Wiley, New York, 1979,2nd
cdn.)
8 NEWTON, G.C., Jr., GOULO, L.A., and KAISER, J.F.: ‘Analytical
design of linear feedback controls’ (Wiley, New York, 1957), pp. 367-
381
9 BYE, C.T., IIARTMANN, G.L., and KILLEN, A.: ‘Development of a
FOG-based GPS/INS’. Proceedings of IEEE Position location and
navigation symposium, 1098, pp. 264-271
10 COX, D.B.: ‘Integration of GPS with inertial navigation systems’ in
‘Global positioning system‘, Vol. I, Institute of Navigation, 1980, pp.
144-153
II GELB, A.: ‘Applied optimal estimation’ (MIT Press, MA, 1974)
12 BROWN, K.G., and HWANG, PY.C.: ‘Introduction lo random signals
and applied Kalman filtering’ (Wiley, New York, 1997, 3rd e h . )
IEE Proc.-Radac Sonar Navrg., Vol. 14X, No. 4, Augu:urt 2001