IEEE - Aerospace and Electronic Systems - April 2022 - 34

Multichannel SDR Platforms for Mobile PCL Applications
Figure 8.
Block diagram of training signal injection circuit.
internal architecture of the USRP, and the interchannel
phase shifts vary by less than 0.15, according to the results
of our experiments. However, due to substantial weight,
power consumption, and host machine requirements (desktop
or server machine, preferable with Intel Core-i7 CPU
or equivalent), this configuration may be installed only in
the case oflarger moving platforms.
LIGHTWEIGHT SYSTEM
The lightweight version of the signal acquisition system is
based on two RSPduo receivers [14]. These receivers can
acquire RF signals with carrier frequencies up to 2 GHz,
with 14-bit ADC performance at 2 MHz sample rate, and
using low-cost USB 2.0 connections for both power supply
and data transfer.
Possible applications of RSPduo devices for multichannel
RF signal acquisition in small-sized systems have
been demonstrated in [19] together with their limitations.
It was pointed out there that due to lack of a common trigger
signal, the coherent acquisition becomes a challenge.
To solve this problem, a method to estimate phase-shifts
and time-delays introduced in the hardware has been
developed and implemented [20] that is based on
matched-filtering of signals received with superimposed
pilot sequences. These sequences are generated in additional
hardware parts and coupled with the signals from
the antennas. This method has been evaluated on real signals
acquired during APART-GAS (Active Passive Radar
Trials - Ground-based, Airborne, Sea-borne) NATO campaign
in September 2019 by the working group SET-242
Figure 9.
Four-channel RF signal acquisition system with training signal
injection.
(Passive Coherent Locators on Mobile Platforms). The
signals were acquired in a four-channel experimental system
designed for PCL and installed on a sea-borne moving
platform (a navy ship).
In order to perform cross-correlation computations, a
training sequence with good correlation properties (pseudorandom
sequence with BPSK modulation) needs to be periodically
injected to the RF signals coming from the antennas.
These signals can be combined by using a power
divider and a number of directional couplers connected to
the inputs ofRSPduo devices, as shown in Figure 8.
The details of the training signal properties, generation,
and processing have been disclosed in [20]. The
proof-of-concept implementation of this method (without
a host machine, which was a typical notebook with Intel
Core-i3 CPU) has been presented in Figure 9.
After compensating phase shifts and time delays offline
in the MATLAB environment, the signals from the reference
channel x4[n] and one of the receiving channels x1[n]
(both derived from two RSPduo devices) have been processed
in the PCL algorithm, shown in Figure 10. The pipeline
is composed of a clutter-removal filter (based on a
lattice structure described in detail in [21-24], [26], [27])
and a cross-ambiguity function calculator described in [25]
and [27]. The final product ofthese algorithms is the series
of 2-D plots S[v,R] showing signals from detected objects
in the bistatic velocity/range (v/R) plane. The remaining
channels x2[n], x3[n] are intended to be processed in a more
advanced version ofthe PCL algorithm.
Figure 10.
Simplified block diagram ofPCL processing algorithm.
34
IEEE A&E SYSTEMS MAGAZINE
APRIL 2022
IEEE - Aerospace and Electronic Systems - April 2022

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