Signal Processing - September 2017 - 40

Correct Frequency Window
Example of a
Correct
Acquisition Mesh

0.15
0.1
0.05
0

0

00

0
50

0
0
50
00
,0
-1

4 2
Code Phase 0 -2
Delay (Chips)

1,

6

Frequency Error (Hz)

Received
Signal
Acquisition Block

Signal
Acquired?

Unambiguous Acquisition
Processing (Optional)

Yes

Incorrect Frequency Window

Example of a
Noisy Mesh
(Signal Not
Acquired)

× 10-4
1.2
1
0.8
0.6
0.4
0.2
0

Tracking
Block

No

2
1, ,00
4 2
5
0
1
0
50 ,00 00
-2 0
Code Phase
0
0
Delay (Chips)
Frequency Error (Hz)
6

Unambiguous Tracking
Processing (Optional)

Continue Searching
Via Changing
Time-Frequency
Window or Code

Reference
Code
Generator

Code Delay and
Code Phase

FIGURE 2. The acquisition-tracking chain of a GNSS signal, with the possible places of the unambiguous processing.

processing stages, as seen in Figure 2, can be added in the
acquisition stage, the tracking stage, or both stages (see the
section "Unambiguous Solutions"). We can classify the code
trackers according to three classes, shown in Figure 3: 1)
code trackers based only on correlation outputs and typically
relying on at least three correlators, 2) code trackers based on
some form of subspace processing, and 3) code trackers
based on other forms of nonlinear processing. In addition to
the code-tracking structures that output a code delay estimate, there are also alternative structures in GNSSs, which
compute directly the position/velocity/time solution [10].
The class corresponding to multicorrelator structures [11],
[12], which is also the class used in the simulations shown in
this article, includes the majority of the code trackers. Typically, multicorrelator structures are quite robust to noise and
those with more than three correlators are also usually able to
cope with multipaths, to some extent. Multipaths refer to the
nonline-of-sight (NLOS) components due to signal reflections, which affect the accuracy of the line-of-sight (LOS)
delay estimate if unmitigated. The two most known and
40

widely used code trackers in this category are the narrow correlator (NCORR) and the high-resolution correlator (HRC)
[13], which are the ones used in our simulations. The second
category of code trackers, based on subspace decomposition, includes algorithms well known in the signal processing community, such as multiple signal classification or space
alternating generalized expectation maximization [12], [14].
Usually, such approaches have good accuracy and multipath
mitigation capability in very good signal conditions, but they
are sensitive to noise. A third category of code trackers, as
shown in Figure 3, includes various nonlinear processing
algorithms, such as peak tracking and the wavelet transform
[2], [11]. Typically, such algorithms enhance the performance
in multipath, at the expense of a higher complexity or less
robustness to noise than other categories of code trackers. A
direct estimation of the receiver position, without explicitly
obtaining the delay estimates, is also possible and is done
in the alternative structures shown on the right-hand side of
Figure 3. Examples in this category includes the vector delay
locked loop and the direct position estimator (DPE) [10].

IEEE SIGNAL PROCESSING MAGAZINE

|

September 2017

|
Table of Contents for the Digital Edition of Signal Processing - September 2017
