International Journal of Wireless Information Networks (2024) 31:155–162

https://doi.org/10.1007/s10776-024-00619-z

Intelligent Reflecting Surface Framework for ED Based Spectrum

Sensing
Rahul Kumar1 · Shweta Singh2

Received: 9 July 2023 / Revised: 9 July 2023 / Accepted: 14 January 2024 / Published online: 20 February 2024
© The Author(s) 2024, corrected publication 2024

Abstract
According to a recent study, adding intelligent reflecting surfaces (IRS) to propagation environments improves the perfor-
mance of wireless devices. The capabilities of cognitive radio’s spectrum sensing over environments with assisted wireless
propagation via IRS are examined in this research. Taking channel fading effects into consideration, we derive closed-form
analytical expressions for the average probability of detection (APD) for single user and cooperative users. Further, the
derived APD expression is used to derive the average area under the receiver operating characteristics curve ( AUC ) expres-
sion. Monte Carlo simulations are also carried out in order to validate the derived expressions. Results demonstarte the
extent of detectability performance in cognitive radio on utilizing the IRS aided wireless propagation environment therefore
witnessing its significance for upcoming wireless technological advancement by offering spectrum efficiency with good
quality of service (QoS).

Keywords Cognitive radio · IRS · Energy detection · Spectrum sensing · Fading

1 Introduction including energy detector (ED) [1–3], eigenvalue-based

detector [4], matched-filter detector [5] and cyclostationary
The wireless industry is looking towards technological feature detector [6]. Amongst all ED has been widely studied
advancements like 5G, 6G, and beyond services with an due to its low computational complexity, and fast sensing
aim to enhance the communication capacity and quality of performance that is a crucial parameter of SS where there is
service (QoS) to the users. In response to such rapidly rising rapid transition from one band to another. However, ED per-
demand, we need a system that offers a spectral and energy formance is constrained by channel impairment issues like
efficient model while fulfilling the higher rate requirement. shadowing, multipath reflections, fading, and path losses,
Cognitive Radio (CR) is one such system that has been that makes the PU signals at the SU so weak for accurate
widely studied to offer the spectral efficiency, however due sensing. To overcome this limitation, recent researches have
to unpredictable nature of propagation channel behaviour reported the use of digitally controlled and dynamically-
and its unavoidable losses, leads to inaccurate detection per- tunable meta-surface to improve the network performance.
formance of CR system that restricts its usage due to nega- Such smart intelligent reflecting surface (IRS) elements
tive interference to primary user (PU) and/or low spectral adopt joint phase control and radio propagation environ-
efficiency (SE) for secondary user (SU). Numerous studies ment control to enhance received signal performance [7].
have been conducted on the design of various detectors and Due to this appealing feature IRS, it has recently emerged as
the algorithms for the spectrum sensing (SS) of CR system crucial enabler technology to improve signal coverage and
enhance spectral efficiency. Additionally, the deployment
of IRS in CR network can meet the demand of maximizing
* Shweta Singh
[email protected] SU throughput and mitigate SU interference level at PU [8].
Motivated by the above features, IRS assisted CR-net-
1
Department of Electronics and Communication Engineering, work has recently drawn much research interest [9] [10].
GLA University, Mathura, U.P., India Where, IRS exploits the large aperture and passive beam-
2
Department of Electronics and Communication Engineering, forming gains to boost the PU signal strength received at the
Manipal Institute of Technology Bengaluru, Manipal SU. This configuration facilitates the SS in a cost-effective
Academy of Higher Education, Manipal, India

Vol.:(0123456789)
156 International Journal of Wireless Information Networks (2024) 31:155–162

manner and also offers better spectrum sharing [11] and • Secondly, in order to get better understandings on CR oper-
improved spectrum sensing [12] performance. Other related ating behaviour we derive closed form analytical expres-
work based on IRS assisted CR network assumes optimiza- sion for average area under the receiver operating charac-
tion problem for evaluation of performance metrics of the teristics ( AUC).
aforementioned system. In [8] IRS is employed to improve • Based on derived performance parameters we analyses
the SE of the SU. The formulated problem is non-convex ED detection performance for different number of IRS
in nature and is solved by alternating optimization (AO) elements, distance between nodes and time bandwidths etc.
algorithm. AO is further studied in [13] to solve another
such nonconvex problem of maximizing SE of the SU by Rest of the article is organised as;
joint optimizing the transmit precoding matrices at SU, the Section 2 provided the proposed system model for IRS
phase shift matrix at IRS, and power-splitting factors at the assisted ED based spectrum sensing and SNR statistics. Deri-
user. Further, another such optimization problem is dealt in vation of various performance measures is provided in Sect. 3.
[14] by adopting inexact block coordinate descent (IBCD) Related discussions on ED performance through numerical
optimization algorithm. where the phase matrix at the IRS, simulation have been presented in Sect. 4. In the end conclu-
the precoding matrix, and the artificial noise covariance at sion to important findings in this article are provided in Sect. 5.
the SU are optimally designed to achieve maximal secrecy
performance of the SU assuming imperfect knowledge of the 2 System Model and SNR Characteristics
PU channel state information (CSI). Such intractable non-
convex optimisation problem constitutes a several design 2.1 System Model
challenges and requires computationally intensive solution.
This complexity in design is further escalated by increase As illustrated in Fig. 1, the conceptual framework of the pro-
in number of IRS elements and hence poses constraint in posed system comprises a primary user (PU), secondary user
practical realization of massive scale network. The authors (SU), and an intelligent reflecting surface (IRS) that is made
in [15] have carried out statistical analysis and evaluated the up of N reflective elements. The SU is tasked with detecting
average detection probability of IRS assisted CR network but the existence of an active signal transmitted by the PU, at the
the study is limited to only detection probability evaluation presence of the IRS. The signal that the SU receives through
and not considered the average area under curve (AUC) per- the IRS can be represented as follows.
formance parameter. However, the above research, assumes {
that the SU has statistical/ instantaneous CSI, which is prac- n ∶ H0
y=
𝛼.x + n ∶ H1 (1)
tically difficult to attain. Since, the PU position is random
and time variant due to which the channel is uncontrollable
where x, 𝛼 and n represents transmitted PU signal, PU-
and random hence assuming statistical CSI information is
IRS-SU channel gain and additive white gaussian noise
impractical.
respectively. The presence (i.e. hypothesis H1) or absence
Motivated, by the above literature study where the prior
(hypothesis H0 ) of PU signal is decided by comparing deci-
studies have mainly focussed either on computationally
sion threshold 𝜉 with the energy ( Ξ ) of received signal y.
intensive solutions or assumption of CSI which practically
The two performance probability are defined as probability
waves-off the potential effect of simplicity in design of pas-
sive elements of IRS based system on digital communica-
tion system performances and a very few have considered
the statistical analysis of the detection performance of CR
system. Therefore, with a focus on the significance of the
issue, this article investigates the ED detection performance
in an IRS aided propagation environment. Considering large
number of IRS elements and leveraging central limit theo-
rem to characterise PU-IRS-SU SNR statistics we derive
various performance measures for the ED based spectrum
sensing. The key contributions to this work are summarised
as follows:

• Firstly, spectrum sensing performance for the proposed

model is examined via deriving closed form expressions
for average probability of detection (APD) for single user
and cooperative user scenario. Fig. 1  System model
International Journal of Wireless Information Networks (2024) 31:155–162 157

( )
of detection
( as P)r Ξ > 𝜉|H1 and probability of false alarm respectively [17]. Therefore R2 will be a non-central chi-square
as Pr Ξ > 𝜉|H0 . Considering non-faded channel conditions RV with one degree of freedom and PDF as given by [18].
(i.e. 𝛼 constant), the two probabilities are as follow [16], �√ �
� �−1∕4 r+𝜆clt
�√ √ � 1 r − 2 r𝜆 clt
pd ( 𝛾�𝜉) = Qd 2𝛾, 𝜉 (2)
fR2 (r) = 2 e 2𝜎clt I−0.5 2 (7)
2𝜎clt 𝜆clt 𝜎clt

and where, In (.)represents modified Bessel function first kind

( ) with nth order,𝜆clt = (N𝜋∕4)2and 𝜎clt2
= N(1 − 𝜋 2 ∕16). Now
1 𝜉
pfa (𝜉) = Γ d, (3) from (6) the PDF of instantaneous snr is expressed as:
Γ(d) 2
� � �−1∕4 �√ �
𝜆 𝛾̄ +𝛾d� �𝜆
d� d − clt 2 𝛾d
where, Qd (., .) is Marcum-Q function, Γ(.) is Gamma func- f𝛾 (𝛾) = 2 𝛾 −0.25 e 2̄
𝛾 𝜎
clt I
−0.5 √ 2
clt

tion, and Γ(., .) is upper incomplete Gamma function, d is the 2𝜎clt 𝛾̄ 𝛾̄ 𝜆clt 𝛾̄ 𝜎clt
time-bandwidth product and 𝛾 represents instantaneous (8)
E
signal-to-noise ratio (SNR) defined as, 𝛾 = |𝛼|2 Np with Ep A more simplified expression for PDF in (8) can be obtained
0
and N0 representing PU transmitted power and noise power by expanding Bessel function as summation series i.e.
spectral density respectively. ∞
( )p
𝜆
− clt ∑ 1 𝜆clt
2𝜎 2
f𝛾 (𝛾) = e clt 2
𝛽 p+0.5 𝛾 p−0.5 e− 𝛽 𝛾
p!Γ(p + 0.5) 2𝜎
2.2 SNR Characteristics p=0 clt
(9)
d�
For the given propagation framework, PU-IRS-SU channel where, 𝛽 = 2 .
2̄𝛾 𝜎clt
Observe (9) consist of ∞ series and related
coefficient can be written as follow: series truncation terms i.e. P can be obtained by using PDF
area property ∫ f𝛾 (𝛾)d𝛾 = 1 , as given by (11).
∞
N
∑
𝛼= hi ej𝜙i gi (4) 0
( )p
P
i=1 𝜆
− clt ∑ 1 𝜆clt
2𝜎 2
f𝛾 (𝛾) = e clt 𝛽 p+0.5 𝛾 p−0.5 e− 𝛽 𝛾
where, N represents number of IRS elements, 𝜙i is the phase p!Γ(p + 0.5) 2
p=0 2𝜎 clt
shift induced by ith IRS element. hi and gi denotes channel (10)
gains between PU to ith IRS element and ith IRS element to P p
∑ 𝜆clt −
𝜆clt

SU such that hi = dpr 𝛽i e−j𝜓i and gi = drs 𝜀i e−j𝜑i . Where

−𝜍∕2 −𝜍∕2 2𝜎 2
2 p
e clt ≈1 (11)
dpr and drs are link distances between PU to IRS and IRS p=0 p!(2𝜎clt )
to SU respectively, 𝜍 is path loss exponent, 𝜑i and 𝜓i are
associated phases corresponds to channel gains hi and gi Table 1 shows, the minimum number of truncation terms
respectively. (P) for different IRS elements (N) required to satisfy condi-
From [17] maximised SNR at SU can be attained by set- tion (11).
ting up 𝜙i = 𝜑i + 𝜓i . Therefore from (1)
3 Performance Analysis
N
1 �
𝛼=√ 𝛽i 𝜀i (5) 3.1 Detection Probability at Stand‑Alone SU
d� i=1

where d� = dpr
𝜍 𝜍 .Hence received instantaneous snr at SU
× drs Detection probability expressed in (3) considers no fading
terminal can be written as; effects although practical propagation scenario are subjected
to fading, therefore channel coefficient is no longer constant,
(N )2
but rather fluctuate randomly. Thus, the determination of the
1 ∑ EP 1
𝛾= � 𝛽𝜀 = � R2 𝛾̄ (6) average detection probability can be expressed in the follow:
d i=1 i i N0 d

where 𝛾̄ denotes average SNR defined as 𝛾̄ = EP ∕N0 and

N
∑
R = 𝛽i 𝜀i.
i=1
For a significant number of IRS elements, specifically when
N >> 1 and in accordance with the central limit theorem, the Table 1  Truncation terms for N 10 20 30 50
PDF of R can be approximated by Gaussian random distri- different L values
bution with mean and variance as N𝜋∕4 and N(1 − (𝜋∕4)2 ) P 21 34 45 67
158 International Journal of Wireless Information Networks (2024) 31:155–162

∞
�√ √ � 3.2 Detection Probability with Collaborative SUs
∫ (12)
p̄ d (𝜉) = Qd 2𝛾, 𝜉 f𝛾 (𝛾)d𝛾
0 Numerous research studies have demonstrated that energy
detection-based spectrum sensing can be substantially
where, f𝛾 (𝛾) represents PDF of instantaneous SNR in fading enhanced through the cooperation among the of SU by
environment. sharing detection information. When multiple users col-
Using integral form of Marcum-Q function [19, eq.(1)], laborate, the probability of detection and false alarm can
in (12) followed by [20, Eq. (8.445)], (12) can be repre- b e e x p r e s s e d a s p̄ css
Δ
= 1 − (1 − p̄ d (𝜉))Hc a n d
d
sented as: p̄ css
Δ
= 1 − (1 − pfa (𝜉))Hc respectively [21]. Utilizing these
fa
∞
∑
∞ expressions, the average probability of detection for a col-
Υ(d + s, 𝜉∕2)
s!Γ(d + s) ∫
p̄ d (𝜉) = 1 − e−𝛾 𝛾 s f𝛾 (𝛾)d𝛾 (13) laborative scenario involving Hc users for the given propa-
s=0
0
gation scenario is derived as follows:
(
with (13) and (10) and [20, eq.(6.643.2 )], the average detec-
∞ P
Δ
𝜆
− clt ∑ ∑ Υ(d + s, 𝜉∕2) Γ(s + p + 0.5)
2𝜎 2
tion probability is obtained in (14). p̄ css
d
= 1 − e clt

s=0 p=0
s!Γ(d + s) p!Γ(p + 0.5)
𝜆clt ∞ P
∑ ∑ Υ(d + s, 𝜉∕2) Γ(s + p + 0.5) ( )p )Hc
−
p̄ d (𝜉) = 1 − e 2𝜎 2
clt
𝜆clt 𝛽 p+0.5
s=0 p=0
s!Γ(d + s) p!Γ(p + 0.5) 2
2𝜎clt (1 + 𝛽)s+p+0.5
( )p
𝜆clt 𝛽 p+0.5 (16)
2 ( ( ))Hc
2𝜎clt (1 + 𝛽)s+p+0.5 Δ 1 𝜉
p̄ css =1− 1− Γ d,
(14) fa Γ(d) 2
( ) Hc (17)
Δ 1 𝜉
=1− Υ d,
Γ(d)Hc 2
3.1.1 Proposition A

The truncation error, due to limiting infinite series in (14)

to S finite terms, can be expressed as in (15). Where the 3.3 AUC Performance
value of truncation term S depends on the number of IRS
elements, and SNR, etc., and it is chosen in such a way that Area under ROC is another performance metric used to
the error is confined within a certain limit. assess an energy detector’s effectiveness and have its util-
ity when comparing two systems. AUC computation is
|E| ≤ e
P
Υ(d + S + 1, 𝜉∕2) ∑ Γ(S + p + 1.5)
𝜆clt
−
2𝜎 2
clt
independent of decision threshold hence can serve as a
Γ(S)Γ(d + S + 1) p=0 p!Γ(p + 0.5) single figure to gain better insights on ED receiver perfor-
( )p ( ) mance. Computation of AUC is as follow:
𝜆clt 𝛽 p+0.5 1
2 2 F1 1, S + p + 1.5;S; ∞
2𝜎clt (1 + 𝛽)p+S+1.5 1+𝛽 𝜕pfa (𝜉)
∫ (18)
AUC = − p̄ d (𝜉) d𝜉
(15) 𝜕𝜉
0
Table 2 lists the number of terms S needed to ensure that the
error |E| ≤ 10−7 with pfs = 0.1, d = 10, dpr = drs = 5,𝜍 = 2.3 where,
and different values of SNR and IRS elements.
𝜕pfa (𝜉) 𝜉 d−1 e−𝜉∕2
=− (19)
Proof See Appendix. A  ◻ 𝜕𝜉 2d Γ(d)

From (18) ,(19) and (14) we get,

Table 2  Truncation terms S for |E| ≤ 10−7

N =10 N =20 N =30

SNR=-5 dB 3 4 5
SNR=0 dB 4 5 6
SNR=+5 dB 5 8 9
International Journal of Wireless Information Networks (2024) 31:155–162 159

−
𝜆clt
spectrum sensing as the number of IRS elements increases,
2𝜎 2
AUC = 1 − e clt
at both 0 dB and 5 dB SNR levels. This can be attributed
( )p
∑∞ P
∑ Γ(s + p + 0.5) 𝜆clt 𝛽 p+0.5 to the fact that an increase in the number of IRS elements
results in better beam-forming at the IRS, leading to a
(20)
2
s!p!Γ(p + 0.5) 2𝜎clt (1 + 𝛽)s+p+0.5
s=0 p=0 higher SNR at the SU and thereby improving the ED sens-
∞
Υ(d + s, 𝜉∕2) 𝜉 d−1 e−𝜉∕2 ing performance. Additionally, it is noteworthy that even at
∫
d𝜉 lower SNRs, the performance of the ED detection system is
Γ(d + s) 2d Γ(d)
0 improved with larger numbers of IRS elements. For instance,
when N=10 (SNR=+5dB) and N=20 (SNR=0dB), and
with the aid of [20, eq.(6.445.2)] AUC is given as: similarly for N=20 (SNR=+5dB) and N=40 (SNR=0dB),
− clt ∞ P 𝜆 improved ED performance is observed for lower SNRs (due
2 ∑ ∑ Γ(2d+s) Γ(s+p+0.5)
AUC = 1 − e 2𝜎clt Γ(d)Γ(d+s+1) s!p!Γ(p+0.5) to the presence of a larger number of IRS elements) indi-
cating efficacy of IRS-assisted framework detecting low-
s=0 p=0
� �p � �
𝜆 𝛽 p+0.5 1
× 2𝜎clt2 2 2d+s
clt
(1+𝛽) s+p+0.5 2 F1 1, 2d + s, d + s + 1,
2 powered PUs.
(21) Figure 3 illustrates the effects of PU-IRS-SU link dis-
tances on the detection performance as a function of the
average received SNR. The results demonstrate a signifi-
4 Results and Discussion cant degradation in the ED detection performance when the
PU-SU link distance is increased. However, an improvement
This section presents an analysis of the performance of ED in the detection performance is observed with an increase
based spectrum sensing for the introduced IRS assisted sys- in the IRS elements (N), for instance, the POD increases
tem model. The aim is to demonstrate the efficacy of the ED by 25.4% and 67.88% when N increased to 20 and 30 (from
based spectrum sensing, and this is achieved through the its initial value of 10 at dpr = drs = 10 and SNR=+10dB),
use of various curves, including ROC ( p̄ d vs pfa ), p̄ d vs 𝛾̄ , similarly, an improvement of 142.2% and 234.7% is observed
and AUC . Drawn curves are obtained through the derived (at dpr = drs = 5) indicating that a large aperture IRS can
analytical expressions and verified with the corresponding effectively sense PU with low SNR.
Monte Carlo simulations. Figure 4 depicts the p̄ d versus SNR curve for the coopera-
Figure 2 presents the ED detection performance using tive spectrum sensing scenario. The results reveal that the
the ROC curve for varying numbers of IRS elements (N) detection capabilities are further improved with the incor-
and SNR levels. The results obtained demonstrate a signifi- poration of cooperation between the SUs. It can be observed
cant improvement in the detection capability of ED based that the POD improves by 59.83% and 89.70% (with N=30,
SNR=0dB) when considering 2 and 3 cooperative users,
respectively. Furthermore, the observed improvement is

1
1
0.9
N=10
0.9 N=20
0.8
N=30
Simulations
0.7 0.8
d pr=drs =5
0.6 0.7

0.5 0.6
N=10, SNR=0dB
0.4 N=20, SNR=0dB
0.5
N=40, SNR=0dB
0.3 N=60, SNR=0dB
N=10, SNR=5dB 0.4
0.2 N=20, SNR=5dB d pr=drs =10
N=40, SNR=5dB 0.3
0.1 N=60, SNR=5dB
Simulations
0.2
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0.1
-20 -15 -10 -5 0 5 10 15 20 25

Fig. 2  Average probability of detection versus Probability of false

alarm (ROC) with d = 1,𝜍 = 2.1, dpr = 5 ,drs = 4 and different num- Fig. 3  Pd versus SNR (dB) Plots, with d = 1, 𝜍 = 2.1, pfa = 0.1 dif-
ber of IRS elements (N) and SNR ferent number of IRS elements (N) and link distances (i.e. dpr & drs)
160 International Journal of Wireless Information Networks (2024) 31:155–162

1
SNR=+15 dB, almost equal AUC performance is observed,
0.9 indicating an important outcome that a 10 dB SNR require-
ment can be met by simply increasing the number of IRS
0.8
elements. Additionally, an interesting observation is that
0.7 poor detection performance is observed when the PU or SU
is at an equal distance from the IRS. In contrast, for unequal
0.6
distances from the IRS, the performance is comparatively
0.5 better.
0.4
Figure 6 presents the AUC vs d curve for different IRS
elements and path loss conditions. It is evident that under
0.3 the given path loss condition, the time-bandwidth product
0.2
d plays a critical role in determining detector performance.
As the value of d increases, the AUC performance reduces
0.1
-15 -10 -5 0 5 10 15 20 25
since both APD and PFA increase with an increase in d, but
PFA increases at a higher rate leading to an overall reduc-
tion in AUC . Additionally, the behaviour is studied at 𝜍 =2
Fig. 4  Pd versus SNR(dB) Plots for cooperative detection, with d = 1, and 𝜍 =3, which demonstrates the worst receiver detection
𝜍 = 2.1, pfa = 0.1, dpr = 10 & drs = 10 and different number of IRS performance at 𝜍 =3, as the AUC lies near 0.5. However,
elements (N)
its performance improves significantly when the number of
IRS elements is increased. Conversely, for 𝜍 =2, a superior
200.79% and 298.82% at 30 dB SNR. Hence, the utilization receiver performance is observed due to better path loss
of cooperation among SU, along with a larger size IRS, has conditions, which further improves with an increase in the
the potential to sense a PU located at larger distances or with number of IRS elements.
low transmitted power.
Figure 5 portrays the ED detection performance in terms
of the AUC characteristics with different dpr values, where 5 Conclusion
dpr + drp = 15, at SNRs levels of 5 dB and 15 dB. It can be
observed that at lower SNRs and with fewer IRS elements The presented research investigates the performance of
(i.e. SNR=+5 dB, N=10), the receiver performs poorly, and ED spectrum sensing over of intelligent reflecting surfaces
the AUC lies close to its lower bound, i.e., 0.5. However, for (IRS) assisted communication systems. To this end, Novel
higher SNRs and with more IRS elements (i.e. SNR=+15 expressions for two performance measures i.e. average prob-
dB and N=20 or 30), an improved AUC performance is ability of detection with (i) single user and (ii) cooperative
observed. Moreover, at N=30, SNR=+5 dB, and N=10,

1 1

0.95 0.95

0.9 0.9 =2

0.85 0.85

0.8 0.8
SNR=+15dB
0.75 0.75

0.7 0.7

0.65 0.65

0.6 SNR=+5dB 0.6

=3
0.55 0.55

0.5 0.5
2 4 6 8 10 12 14 5 10 15 20 25 30 35 40 45 50 55 60

Fig. 5  AUC versus dpr (dpr + drs = 15) with d = 1, 𝜍 = 2.5 and differ- Fig. 6  AUC versus time-bandwidth product (d) with dpr = 5, drs = 5,
ent number of IRS elements (N) and SNR SNR=10 dB and different number of IRS elements (N) and 𝜍
International Journal of Wireless Information Networks (2024) 31:155–162 161

detection and average area under the ROC curve are derived. need to obtain permission directly from the copyright holder. To view a
The results demonstrate an improved detection performance copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/.
of ED-based spectrum sensing when number of IRS ele-
ments are increased which can be further improved by incor-
poration of cooperation between SUs. Hence enhancing ED References
capabilities to identify the presence of low-powered PU or
1. A. Ali and W. Hamouda, Advances on spectrum sensing for cogni-
operating over larger distances. Therefore, using IRS has tive radio networks: Theory and applications, IEEE Communica-
the potential to be a highly effective solution for spectrum tions Surveys & Tutorials., Vol. 19, No. 2, pp. 1277–304, 2016.
sensing in future cognitive radio-based wireless communi- 2. I. Sobron, P. S. Diniz, W. A. Martins and M. Velez, Energy detec-
cation systems. tion technique for adaptive spectrum sensing, IEEE Transactions on
Communications, Vol. 63, No. 3, pp. 617–27, 2015.
3. A. Nasser, Al Haj Hassan H, J. Abou Chaaya, A. Mansour and K.
C. Yao, Spectrum sensing for cognitive radio: Recent advances and
Proof: Proposition A future challenge, Sensors, Vol. 21, No. 7, pp. 2408, 2021.
4. T. Yucek and H. Arslan, A survey of spectrum sensing algorithms
for cognitive radio applications, IEEE Communications Surveys &
The expression for the truncation error resulting from the Tutorials., Vol. 11, No. 1, pp. 116–30, 2009.
truncation of an infinite series into a finite number of terms, 5. X. Zhang, R. Chai, and F. Gao, Matched filter based spectrum
can be expressed as: sensing and power level detection for cognitive radio network. In
2014 IEEE global conference on signal and information processing
P
( )p (GlobalSIP) 2014 Dec 3 (pp. 1267-1270). IEEE.
𝜆
− clt ∑ 1 𝜆 𝛽 p+0.5
2
|E| = e 2𝜎clt clt 6. G. Huang and J. K. Tugnait, On cyclostationarity based spectrum
p!Γ(p + 0.5) 2𝜎clt 2
(1 + 𝛽)p+0.5 sensing under uncertain Gaussian noise, IEEE Transactions on Sig-
p=0
∞
(22) nal Processing, Vol. 61, No. 8, pp. 2042–54, 2013.
∑ Υ(d + s, 𝜉∕2) Γ(s + p + 0.5) 7. M. Di Renzo, A. Zappone, M. Debbah, M. S. Alouini, C. Yuen, J. De
Rosny and S. Tretyakov, Smart radio environments empowered by
Γ(d + s) s!(1 + 𝛽)s
s=S+1 reconfigurable intelligent surfaces: How it works, state of research,
and the road ahead, IEEE Journal on Selected Areas in Communica-
Observe represents regularised lower incomplete
Υ(d+s,𝜉∕2)
Γ(d+s) tions, Vol. 38, No. 11, pp. 2450–525, 2020.
gamma function and decreases with increase in s , hence 8. J. Wang, Spectral efficiency maximization for IRS-assisted wireless
communication in cognitive radio networks, Physical Communica-
maximising right hand side and reordering series from 0 to tion, Vol. 1, No. 50, 101528, 2022.
∞ we get. 9. X. Tan, Z. Sun, J. M. Jornet, and D. Pados, Increasing indoor spec-
trum sharing capacity using smart reflect-array. In 2016 IEEE Inter-

|E| ≤ e
P
Υ(d + S + 1, 𝜉∕2) ∑ national Conference on Communications (ICC) 2016 May 22 (pp.
𝜆clt
− 1
2𝜎 2
clt
1-6). IEEE.
Γ(d + S + 1) p=0 p!Γ(p + 0.5)
10. L. Zhang, Y. Wang, W. Tao, Z. Jia, T. Song and C. Pan, Intelligent
( )p ∞
(23) reflecting surface aided MIMO cognitive radio systems, IEEE Trans-
𝜆clt 𝛽 p+0.5 ∑ Γ(t + S + p + 1.5) actions on Vehicular Technology, Vol. 69, No. 10, pp. 11445–57,
2
(1 + 𝛽)p+S+1.5
t 2020.
2𝜎clt t=0 (t + S + 1)!(1 + 𝛽)
11. X. Guan, Q. Wu and R. Zhang, Joint power control and passive
beamforming in IRS-assisted spectrum sharing, IEEE Communica-
Again, rearranging infinite series in (23) as hyper geometric tions Letters, Vol. 24, No. 7, pp. 1553–7, 2020.
function, expression for truncation error bound is given by 12. X. Li, Q. Zhu, and Y. Wang, IRS-assisted crowd spectrum sensing
(15) in B5G cellular IoT networks. In 2020 International Conference
on Wireless Communications and Signal Processing (WCSP), pp.
761–765. IEEE, 2022
Funding Open access funding provided by Manipal Academy of 13. P. Van Quyet, and H.H. Kha, Spectral Efficiency Optimization in
Higher Education, Manipal. IRS-Aided Multiuser MIMO SWIPT Cognitive Radio Systems. In
2022 RIVF International Conference on Computing and Commu-
Open Access This article is licensed under a Creative Commons Attri- nication Technologies (RIVF), pp. 203–208. IEEE, 2022
bution 4.0 International License, which permits use, sharing, adapta- 14. V. T. Duy, and H. H. Kha. Secrecy Rate Optimization for IRS-Aided
tion, distribution and reproduction in any medium or format, as long MIMO Cognitive Radio Systems with SWIPT. In 2022 IEEE Ninth
as you give appropriate credit to the original author(s) and the source, International Conference on Communications and Electronics
provide a link to the Creative Commons licence, and indicate if changes (ICCE), pp. 139–144. IEEE, 2022
were made. The images or other third party material in this article are 15. Abbass Nasser, Hussein Al Haj. Hassan, Ali Mansour, Koffi-Clem-
included in the article's Creative Commons licence, unless indicated ent. Yao and Loutfi Nuaymi, Intelligent reflecting surfaces and
otherwise in a credit line to the material. If material is not included in spectrum sensing for cognitive radio networks, IEEE Transactions
the article's Creative Commons licence and your intended use is not on Cognitive Communications and Networking, Vol. 8, No. 3, pp.
permitted by statutory regulation or exceeds the permitted use, you will 1497–1511, 2022.
16. F. F. Digham, M. S. Alouini, and M. K. Simon, On the energy detec-
tion of unknown signals over fading channels. In IEEE International
Conference on Communications. ICC’03, Vol. 5, pp. 3575–3579.
IEEE, 2003
162 International Journal of Wireless Information Networks (2024) 31:155–162

17. E. Basar, M. Di Renzo, J. De Rosny, M. Debbah, M. S. Alouini and 20. I. S. Gradshteyn, I. M. Ryzhik, A. Jeffrey, and D. Zwillinger, Table
R. Zhang, Wireless communications through reconfigurable intel- of Integrals, Series, and Products. Table of Integrals (2007)
ligent surfaces, IEEE Access, Vol. 7, pp. 116753–116773, 2019. 21. A. Ghasemi, and E. S. Sousa, Impact of user collaboration on the
18. J. G. Proakis, and M. Salehi, Digital communications, Vol. 4, pp. performance of sensing-based opportunistic spectrum access. In
593–620. New York: McGraw-hill, 2001 IEEE Vehicular Technology Conference, pp. 1–6. IEEE, 2006
19. A. Nuttall, Some integrals involving the QM function (Corresp.).,
IEEE Transactions on Information Theory, Vol. 21, No. 1, pp. Publisher's Note Springer Nature remains neutral with regard to
95–96, 1975. jurisdictional claims in published maps and institutional affiliations.