Tracking Weak Signals - Fundamentals of Global Positioning System Receivers: A Software Approach - page 273

5 kHz. It is decided that
this loss is acceptable. Thus the C/A code is generated every 20 ms. Every time
only 1 ms of C/A code is generated, it will be used for 20 ms.
However, the local C/A code can also be generated every second. Because
the maximum mismatch can be 100 ns, this method can be applied to all signal
conditions. Especially when the sampling frequency is not an integer, as discussed
in Section 11.13, this method can still track the input signal. Under this condition
the average mismatch time can be 50 ns, which corresponds to a loss of 0.46 dB.
If the computer speed is adequately improved, the C/A code might be gen-
erated every ms to obtain the maximum output. The improvement, however,
in sensitivity is limited only about 0.14 or 0.46 dB as shown in the preced-
ing paragraphs.

the worst loss because the Doppler frequency used is ±

11.3 CARRIER FREQUENCY MEASUREMENT REQUIREMENT

It was discussed in the Section 10.3 that the sensitivity of a GPS receiver is
limited by how accurately navigation data can be decoded. In order to determine
the navigation data correctly a very accurate carrier frequency is needed. In
addition the carrier frequency can also be used to help determine the pseudorange,
which can be considered as carrier-aided processing, and the details will be
discussed in Sections 11.9 to 11.11. Both arguments suggest that an accurate
carrier frequency is needed. The question is how much accuracy is needed.
As discussed in Sections 3.7 and 3.8, the maximum Doppler frequency change
is about 1 Hz per second. If one second of data is processed through the fast
Fourier transform (FFT), the resolution is about 1 Hz, which matches the rate
of change of the Doppler frequency. If longer data, such as 2 seconds, are used
to measure the frequency, the resolution is about 0.5 Hz. But during this time
the carrier frequency may change 2 Hz; thus this resolution and the frequency
change is not matched. In other words, the frequency measured may cover sev-
eral frequency bins, and the accuracy will not improve. Based on this argument
1 second of data is used to calculate the carrier frequency. Better than 1-Hz
accuracy can be obtained by comparing amplitudes of adjacent frequency bins
and extrapolating the result.

11.4 ONE MILLISECOND DATA PROCESSING AND INPUT DATA
SELECTION

Before the discussion of input data selection, two terms will be defined here.

1. Initial C/A code point (or initial point of a C/A code): This refers to the
first digital point of a C/A code, which is obtained from digitizing the first
chip of the code. This chip is the first one of 1,023 chips generated, as
discussedinSection5.6.