IEEE Aerospace and Electronic Systems Magazine - June 2020 - 41

Giusti et al.
Table 4.

Table 5.

Parameters Setting at the Beginning [20]
Drmin

0.3 m

500 MHz

0.35

Parameters Setting at the Beginning [20]
z1

0.15 0.5 120 dB 0.2

The imaging capability is again satisfied while the
detection capability seems to worsen as the AF exceeds z1
(a3 ¼ -1). Nonetheless, a4 attests that the STAP filter Doppler profile is ideal yet. Therefore, the interference presence
seems not to affect the radar detection capability. Because
a ¼ Àg 3 ¼ À0:15 & 2 ¼ 0:2, a is forced to be zero and
as a consequence, the transmitter illuminates the radar scene
with the same bandwidth (500 MHz) within the third sweep
and new system parameters are computed (see Table 7).
In the third cycle, the AF exceeds the minimum one
required (20 dB) suggesting the actuating function (a3) to
reduce the instantaneous bandwidth to compensate the
SINR reduction due to the presence of a strong interference. This is confirmed by a4 that attests an anomaly of
the STAP filter profile (see Figure 7).
This fact means that the new spectrum environment
affects the instantaneous bandwidth used by the system. To
face this issue, the system senses the spectrum to detect the
interference so as the updating rule block increases a certain quantity weight (g 2) related to the resolution and suggests the actuating function to increase the instantaneous
bandwidth so as to exploit the largest portion of the available spectrum. The cycle is then repeated over time in order
to update the memory and adapt in a dynamic way the system parameters to the environmental changes and spectrum
allocation. The system evolution as a function of the interference frequency is reported in Table 8.
It is worth noting that a strong interference appears
within the tenth observation. In this case, the system manages itself in order to transmit within the available bigger
portion of the free spectrum.

AF [dB]
20

NDB

dr ½m

0.3

CONCLUSION
The design of a cognitive multifunction radar is still an
open research field. Most of the literature proposes a CR
architecture to handle a specific task or mission, while
modern radars exhibit a high degree of flexibility and are
designed to handle several tasks at the same time. In this
context, a cognitive architecture may help in managing
the radar limited resources according to some optimally
criteria and in dependence of the environment.
This article fits into this context and proposes a
criterion to design a CR that is able to handle several functionalities and that must operate in a complex and nonstationary environment. The proposed architecture bases its
functioning on the definition of a set of performance indicators and rules, which drive actions to take. Performance
and rules have been defined by taking into account specific
radar functions, namely, target detection and imaging.
Specifically, performance indicators represent a way to
measure the radar operation success while rules, taking
into account the available resources, act on the system
reconfigurable parameters based on the measured success
with respect to the observed scene and the past acquired
experience. Then, rules are necessarily linked to radar
resources, the system memory and the desired performance. In details, rules are the way all these three elements affecting a radar systems (limited resources,
previous failures/successes, and desired performance) are
balanced to select the most appropriate set of waveform
parameters. Rules, as defined in this article, are dynamic.
As a consequence, they may be updated or changed

Figure 6.
STAP performance assessment at n ¼ 1.

JUNE 2020

IEEE A&E SYSTEMS MAGAZINE

IEEE Aerospace and Electronic Systems Magazine - June 2020

Table of Contents for the Digital Edition of IEEE Aerospace and Electronic Systems Magazine - June 2020