Satellite Signal Acquisition, Tracking, and Data Demodulation - Understanding GPS: Principles and Applications - page 239

that are squared and sum med. As noted in Section 5.8, over one coherent integra-
tion period the envelope has a Ricean pdf when the signal is present and a

Rayleigh distribution when the signal is not present. Further noncoherent integra-
tion is typically achieved by summing I 2

I

2

+ Q

2

Q 2 from each of K successive coherent
integration periods. The total resultant dwell time is KT . The sum of I 2

+

Q 2 over the
K periods has a central chi-square pdf with 2 K degrees of freedom when the signal is
not present and a noncentral chi-square pdf with 2 K degrees of freedom when the
signal is present (e.g., [24]). Figure 5.45 assumes that the thresholds for each pair of
( T , K ) values is set to achieve a false alarm probability of 10 −4 .
The curves in Figure 5.45 ignore some practical issues such as cross-correlation
among C/A code signals, which generally require the thresholds to be inflated some-
what to avoid excessive false alarms. As noted in Section 4.3.4, the cross-correlation
levels of the C/A code signals can be as high as -21 dB. This can cause false acquisi-
tions under certain Doppler differences and antenna gain conditions. For example,
a user in a building might receive one C/A code signal through a window with very
little attenuation and a second C/A signal through several floors of a building with
very high attenutation. Discrimination among signals received with great strength
differences can be very difficult for a C/A code receiver, resulting in an occasional
false acquisition. Fortunately, the unwanted SV signal cannot usually be tracked for
long because both the correlation properties and the Doppler change rapidly, result-
ing in loss of lock and a reacquisition process for the GPS receiver. It is important
that the GPS receiver design implement sophisticated C/A code search procedures
that avoid cross-correlation and autocorrelation sidelobe acquisitions. If the
receiver successfully hands over from C/A code to P(Y) code, then this success pro-
vides a built-in assurance of the integrity of the C/A code measurements. However,
if the receiver remains in C/A code, then there is always concern that false tracking
has occurred under such marginal signal conditions.
It is also important to note that significant implementation losses can result
from Doppler frequency errors in the carrier NCO and timing errors in the replica
PRN code, particularly for large values of T . Thus, the curves in Figure 5.45 should
really be viewed as upper bounds on achievable detection probabilities.
The numerous design tradeoffs and solutions that have been adopted within
industry for indoor applications are discussed further in Section 9.4.

+

5.14 Codeless and Semicodeless Processing

Since only the P(Y) code signal is normally broadcast on L2 by the GPS SVs up
through the Block IIR-Ms, this denies direct two-frequency operation to SPS users
when AS is activated by the control segment (see Section 4.3.1). Two-frequency
carrier-phase measurements are highly desirable for surveying applications (see Sec-
tion 8.4) and dual-frequency pseudoranges are needed to accurately compensate for
ionospheric delays in navigation applications (see Section 7.2.4.1). This has moti-
vated the development of techniques to obtain L2 Y code pseudorange and car-
rier-phase measurements without the cryptographic knowledge for full access to
this signal. These techniques are referred to as either codeless or semicodeless pro-
cessing. Codeless techniques only utilize the known 10.23-MHz chip rate of the Y