IEEE - Aerospace and Electronic Systems - September 2022 - 22

An Experimental Analysis of Cyclic and Reference Signals of 4G LTE for TOA Estimation and Positioning in Mobile...
Figure 2.
Block diagram of a 4G LTE baseband signal processor for TOA estimation.
Note that CIR detection in (6) can be generalized to
include the determination ofslot number and symbol number
via a multidimensional search if they have not already
been determined from PSS and SSS. It can also be used as
an additional confirmation of these parameters if required.
One advantage of using CRS is that the signal is transmitted
over the entire downlink bandwidth up to 20 MHz, not
merely the basic bandwidth of 1.4 MHz available to
PSS/SSS.
In a slot, there are up to six CRS sequences that can be
used to generate six CIRs: two on OFDM symbol 0 for AP
0 and 1, two on OFDM symbol 1 for AP 2 and 3, and two
on OFDM symbol 4 for AP 0 and 1 again. In our limited
collection of data, we mostly encountered AP 0 and 1 on
symbols 0 and 4 and rarely AP 2 and 3 on symbols 1. Examples
of CRS in the time and frequency domains and its
detection with experimental data are shown in the
" Example Signals and Signal Acquisition Results " section.
BASEBAND SIGNAL PROCESSOR FOR TOA
ESTIMATION AND POSITIONING
Figure 2 shows the block diagram of a 4G LTE baseband
signal processor suitable for TOA estimation, ranging,
and positioning. Since PSS and SSS are transmitted within
the basic bandwidth of 1.4 MHz, the baseband signal is
first resampled at the nominal rate of 1.92 MHz with the
number of RBs Nrb ¼ 6 for PSS and SSS detection.
Depending on the transmission bandwidth being 3, 5, 10,
15, or 20 MHz, the baseband signal is resampled again to
one of the five higher nominal rates of 3.84, 7.68, 15.36,
23.04, or 30.72 MHz, if needed, with the number of RBs
Nrb ¼ 15, 25, 50, 75, or 100, as shown in Table 1, for
CRS-based channel estimation and TOA estimation. CP
22
detection is optional but if used, it can speed up the time
search for PSS and SSS.
Due to rather frequent onset of deep fading in urban
mobile environments, it is necessary to determine if an
LTE signal, either direct or multipath, is present prior to
making any further processing. The high volatility of signals
calls for an adaptive detection threshold. Like modern
radar systems for detecting targets in clutter where, due to
constant changes in radar, target, and target background
characteristics, a fixed detection threshold is not practical.
As a result, adaptive techniques can be used to maintain a
constant false alarm rate (CFAR) irrespective of the circumstances
[35]. The CFAR method is used for the detection
ofCP, PSS, SSS, and CRS, respectively, in Figure 2.
As shown in (6), IFFT is applied to the FFT-implemented
circular correlation or to the estimated channel transfer
function to produce the CIR, which describes the direct
and multipath signals in terms of their strength versus delay
relative to the start ofthe sliding window. The peak location
ofeither the earliest arrival (above a CFARdetection threshold)
or the strongest arrival can be taken for fine TOA estimation
through interpolation via a quadratic or sinc-function
curve fitting to within a fractional ofa sample [36].
A more elaborate method to estimate the multipath
signal parameters is to apply the matching pursuit (MP)
[37] and orthogonal matching pursuit (OMP) [38], [39]
algorithms to the CIR [13], [14] in the time domain or the
order-recursive least-square MP algorithm [40] to the
channel transfer function [23] in the frequency domain
[29]. Similar methods were used for 4G LTE signals [21],
[22] and GNSS signals [41]-[43]. A combination of
CFAR and OMP in the frequency domain is presented in
[29] for TOA estimation.
TOA estimates are then converted to pseudoranges
after the unknownTOT is removed together with the
IEEE A&E SYSTEMS MAGAZINE
SEPTEMBER 2022

IEEE - Aerospace and Electronic Systems - September 2022

Table of Contents for the Digital Edition of IEEE - Aerospace and Electronic Systems - September 2022