GPS Receiver-Related Subjects - Fundamentals of Global Positioning System Receivers: A Software Approach - page 319

1-ms time mark

2000 2000 2000 2000 2000 2000

(a) Correct sampling frequency

2000 2002 2004 2000 + p

(b) Sampling frequency too high
FIGURE 12.10 Shift of initial phase of C/A code.

Theoretically this measurement is very simple and only two points are needed.
For example, one can obtain the initial C/A codes at the 1st and 1001st ms through
acquisition. If the initial point has shifted to the right, the sampling frequency is
higher than the nominal value. If the initial C/A code measured in the 1st ms is at
2000 and the initial C/A code measured at the 1001st ms has shifted by p points,
the result should be 2000

p . Since from 1st to 1001st ms covers 1 second of
time, there are a total of 5 , 456 , 000

+

+ p samples. Thus, the actual sampling time

t s is

1
5456000 + p

t s = or f s = 5456000 + pHz ( 12 . 1 )

This simple equation can be used to find the sampling frequency. Because
the noise can cause an error in the measurement, this simple method maybe
inadequate.
In the actual determination of the sampling frequency the process is as follows:
One ms of data is used to perform coherent integration and 20 ms for noncoherent
integration, and the initial point of the C/A code is determined every 20 ms. In 2
seconds this 20-ms operation is performed 100 times. The length of the data used
is not critical but should depend on the signal strength. For a strong signal, less
data can be used. The idea is to use 100 points of 20-ms data averaged to obtain
a more dependable result. A typical initial C/A code measured over 2 seconds of
time is shown in Figure 12.11, which contains 100 data points. Each data point is
obtained from 20 ms of data. Note that most of the data can fit on a straight line,
so we refer to this line as the “desired line,” though there are many obviously
erroneous points. In determining the initial C/A code sometimes the values wrap
around. For example, if in the 1st 20 ms the initial C/A code phase measured
is 4 and the second 20 ms the value is 5454, the first value can be adjusted by
adding 5456. In this example, the first value will become 5460 (4 + 5456). In
general, if there is a wrap around, either the value before the wrap around or the
values after it must be adjusted by 5456.
The following operation is used to eliminate the erroneous points and to find
the slope of the desired line. There might be many ways to accomplish this goal.
One of the methods is presented here. Figure 12.12 shows the difference values