Global Positioning System Reference
In-Depth Information
if the time is offset by
100 Km. The results
here are obtained from YUMA data 139 and shown in Table 12.2.
In this table, column 1 lists the satellite number. The frequencies listed in
column 2 are used as references. Note that for time, latitude, and longitude, the
errors are all zero. Columns 3, 5, 7, and 9 are with different error combinations.
For example, in column 5 the time is off by
±
1 minute, the distance is offset by
±
1 minute, and the latitude and
longitude are both off by 100 Km. Columns 4, 6, 8, and 10 give difference
frequencies with the results in column 1 used as reference.
After the time error and location error are combined, the maximum differ-
ence frequency is about
101 Hz, for satellite 11 in column 6. This maximum
difference frequency is a relatively small value compared to
5kHz.Usingthe
data in both Tables 12.1 and 12.2, one can estimate that the estimated Doppler
frequencies are within ± 150 Hz. If one knows the Doppler frequency of a
satellite in this range, the acquisition can be simplified. Instead of searching
a frequency range of ± 5 kHz, a frequency range of ± 150 Hz should be ade-
quate. The following section discusses acquisition with relative accurately carrier
frequency information.
±
12.3 ACQUISITION WITH ACCURATE CARRIER
FREQUENCY INFORMATION
If the carrier frequency of a certain satellite can be predicted rather accurately,
such as within ± 150 Hz, the acquisition can be rather simple. Not only can
the frequency range search be narrowed, but also the FFT length reduced. The
method of acquisition for a coherent integration of 10 ms will be presented
here.
The input signal will first be down-converted to a baseband frequency. Because
the Doppler frequency is known from the YUMA data calculation, this frequency
can be used in the down-converting operation. The down-converted frequency
will be processed with a bandwidth of ± 50 Hz, because 10 ms of data have
a corresponding frequency resolution of 100 Hz (1/10 ms). The 10 ms of data
(50,000) can be divided into 10 sections. Each section contains 5,000 points, such
as from 1 5000, 5001 10 , 000 ,..., 49 , 5001 50 , 000. The corresponding
points in the 10 sections of data can be summed to the 5000 points shown in
Figure 12.2.
From the resulting 5000 points, circular correlation is performed with 1 ms of
reference C/A code data, and 2500 points are obtained, as shown in Section 7.9.
This operation covers only ± 50 Hz, because the input data used is 10 ms long.
In order to cover
±
150 Hz, the same operation must be performed two more
times using appropriate down-converting frequencies. For example, if the input
frequency estimated through the YUMA data is at f 0 , the input signal can be
down-converted by f 0 , and the operation above can then be performed. To
cover the additional frequency range, the input signal can be down-converted
by f 0 ±
100 Hz, and the same operation also performed two more times. From
Search WWH ::




Custom Search