Aerospace and Electronic Systems Magazine January 2018 - 43

Xu et al.
parameters to be retrieved include aircraft number as well as
motion parameters for each aircraft. It is known that the aircraft
number can be retrieved from the peaks of high-resolution range
profiles (HRRP) [35] for modern wideband radar, where several
gigahertz (GHz) bandwidth can be used to obtain a much higher
range resolution of several centimeters. Nevertheless, the large
bandwidth will increase analog-to-digital (A/D) requirement, data
storage for a RSP subsystem. Moreover, it is known that high
range resolution will also cause difficulties for target detection because the target's backscattering energy will be distributed among
several range units in a single pulse. Besides, the ARC effect will
become more and more obvious among interpulses with the increase of integration time for moving targets. Furthermore, most
existing air search radars are narrowband radar with only several
megahertz (MHz) bandwidth with respect to tens of meters range
resolution, and the aircraft number of MAF cannot be easily obtained from these low-resolution range profiles.
To cope with this problem, a novel MAF signal model [27] is
proposed for narrowband coherent radar in a long TOT. Then, a
quantitative analysis is given on the phase order approximation
of the MAF echoes according to a given set of target motion and
radar system parameters. It is shown that in a coherent integration time duration with a few seconds, multiple scatterers cannot
be easily resolved [27] in a target. It is reasonable to approximate the echoes of a single target as a single component of chirp
signal or cubic phase signal [27] for MAFs with rigid or nonrigid structures. That is, a target is modeled as a point target to
contribute only a single signal component for synthesized MAF
echoes. Subsequently, the third-order polynomial Fourier transform (PFT) [27, 28], i.e., a special case of GFT with the thirdorder polynomial phase compensation, is applied to accurately
estimate the MAF parameters. The parameters retrieved by the
third-order PFT include aircraft number and f1, f2, and f3 for each
aircraft, which are Doppler center, Doppler rate, and Doppler
rate's rate, respectively. The performance of the third-order PFT
is compared with the existing GRT methods in time-frequency
domain as Figure 11 and Figure 12. Based on the parametric
model, the third-order PFT accomplishes the coherent integration in the whole duration of T = 2.56s, while the third-order
polynomial GRT accomplishes the coherent integration in local
window of 0.32 s and noncoherent integration among the outputs
of different windows. It is known that the frequency resolution
is decided by the coherent integration time [27] and the accuracies are decided by the SNR after the integration. Therefore, it
is clearly shown that the third-order PFT has higher resolution
as well as accuracies for f1, f2, and f3 parameter estimation than
the third-order polynomial GRT does. Notably, due to the sharp
peaks generated by the third-order PFT, one dashed curve and
one "+" plus curve are used for two targets as Figure 12, respectively.

BACKGROUND SUPPRESSION
Figure 12.

Normalized peak profiles of different methods. (a) Profiles along f1. (b)
Profiles along f2. (c) Profiles along f3.

JANUARY 2018

Strong clutter and jamming are two main background interferences affecting the effective target detection. Because moving target,
clutter, and jamming have different motions as well as Doppler

IEEE A&E SYSTEMS MAGAZINE

Table of Contents for the Digital Edition of Aerospace and Electronic Systems Magazine January 2018