This article has been accepted for publication in a future issue of this journal, but has not been
fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/LCOMM.2020.3011843, IEEE
Communications Letters
Performance Analysis of Intelligent Reflecting
Surface Aided Communication Systems
Qin Tao, Junwei Wang, and Caijun Zhong
Abstract—This letter presents a detailed performance analysis
of the intelligent reflecting surface (IRS) aided single-input single-
output communication systems, taking into account of the direct
link between the transmitter and receiver. A closed-form upper
bound is derived for the ergodic capacity, and an accurate
approximation is obtained for the outage probability. In addition,
simplified expressions are presented in the asymptotic regime.
Numerical results are provided to validate the correctness of the
theoretical analysis. It is found that increasing the number of
reflecting elements can significantly boost the ergodic capacity
and outage probability performance, and a strong line-of-sight
component is also beneficial. In addition, it is desirable to deploy
the IRS close to the transmitter or receiver, rather than in the
middle.
Index Terms—Intelligent reflecting surface, ergodic capacity,
outage probability, Rician fading
I. I NTRODUCTION
The intelligent reflecting surface (IRS), which can manipu-
late the propagation channel into a favorable shape, has been
regarded as a promising technology for the next generation
wireless communication systems. As such, it has received
considerable interests from both the industry and academia
[1], [2].
Thus far, most of the works on IRS focus on the design
of the phase shift matrix and the transmit beamformer [3]–
[9]. For multiple-input single-output (MISO) systems, the joint
active and passive beamfomer design problem was studied in
[3], and the impact of discrete phase shift was further consid-
ered in [4]. Meanwhile, the energy efficiency of the systems
was characterized in [5]. Later on, a statistical channel state
information (CSI) based design framework was proposed in
[6]. The more general multiple-input multiple-output (MIMO)
system was considered in [7], while the multi-cast MISO
scenario was addressed in [8], [9].
While the aforementioned works have improved our knowl-
edge of IRS-aided communication systems, few works have
studied the analytical performance of IRS aided systems.
In [10], [11], the outage probability and ergodic spectral
efficiency of the MISO systems was studied, assuming that the
channel between the transmitter and IRS is deterministic. In
[12], [13], the outage probability and achievable rate of single-
input single-output (SISO) system was considered, assuming
Manuscript received April 27, 2020, revised June 29, 2020, accepted July
20, 2020. This work was supported in part by the National Key R&D Program
of China 2018YFB1801104 and 2019YFB180003400, the National Natural
Science Foundation of China under Grant 61922071 and 61725104. The editor
coordinating the review of this paper and approving it for publication was G.
Alexandropoulos. (Corresponding author: Caijun Zhong)
Q. Tao, J. Wang, and C. Zhong are with the College of Information Science
and Electronic Engineering, Zhejiang University, Hangzhou, China, and also
with the Zhejiang Provincial Key Laboratory of Information Processing, Com-
munication and Networking, Hangzhou. (email: caijunzhong@zju.edu.cn).
1089-7798 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: UNIVERSITY OF BIRMINGHAM. Downloaded on July 25,2020 at 10:04:11 UTC from IEEE Xplore. Restrictions apply.
IRS
Transmitter
Receiver
Fig. 1. IRS-aided communication systems.
that the direct link between the transmitter and receiver does
not exist.
In practice, the IRS is usually deployed in a position with
line-of-sight (LOS) to both transmitter and receiver, it is
desirable to adopt the Rician fading model. Motivated by this,
in this letter, we present a detailed performance analysis of
IRS-aided SISO systems in Rician fading channels, taking
into account of the direct channel between the transmitter and
receiver. Closed-form expressions are derived for the ergodic
capacity upper bound and outage probability approximation of
the system. In addition, simplified expressions are obtained in
the asymptotic regime. The findings of the letter suggest that
the IRS can provide an effective signal-to-noise ratio (SNR)
gain of N 2 , where N is the number of reflecting elements,
and a diversity order of N + 1 can be achieved.
II. S YSTEM M ODEL
We consider a three-node communication system consisting
of a single antenna transmitter T, a single antenna receiver R
and an IRS with N reflecting elements, as illustrated in Fig. 1.
Since the receiver can overhear the signal from both the
T-IRS-R and T-R links, the received signal y at the R can be
expressed as
√

y = P hT2 Φh1 + g x + n, (1)
where x denotes the normalized transmit signal with unit
energy, g denotes the T-R channel, the N × 1 vectors h1
and h2 denote the T-IRS and IRS-R channels, respectively.
Also, the N × N matrix Φ , diag{θ} denotes the phase shift
matrix, where θ = [ejθ1 , ejθ2 , . . . , ejθN ] with θn ∈ [0, 2π)
representing the phase shift of the n-th IRS reflecting element.
Finally, n represents the additive white Gaussian noise with
zero mean and variance N0 .
Since the IRS is deployed at a position that has LOS to both
T and R, h1 and h2 are modeled by Rician fading. In contrast,
there is no LoS path between T and R, hence, g is modeled
by Rayleigh fading. Therefore, the channel coefficients can be
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/LCOMM.2020.3011843, IEEE
Communications Letters

expressed as with respect to the K1 and K2 , indicating the identical impact

r r ! of the two hop channels.
1 Kl 1
hl = p αl h̄l + h̃l , l = 1, 2,
dl Kl + 1 Kl + 1 B. Outage Probability
1 In this subsection, we analyze the outage probability of the
g = p α3 h̃3 ,
d3 system, which is defined as the probability of the instantaneous
SNR that falls below a pre-defined threshold γth . Mathemati-
where dl and αl , l ∈ {1, 2, 3}, denote the distance and path cally, it is given by
loss exponent of the corresponding channel, while Ki , i ∈
{1, 2} denotes the Rician factor. Also, h̄l , l ∈ {1, 2}, denotes Pout = Prob (γmax ≤ γth ) . (5)
the normalized LoS component, and h̃l , l ∈ {1, 2, 3} denotes
Since the exact distribution of γmax is unknown, we resort
the normalized non-LOS component.
to tight approximations of the outage probability, and we have
Assuming perfect CSI at the IRS, applying the optimal
the following important result:
phase shift matrix Φ give in [14], the maximum SNR of the
Theorem 2: When N → ∞, the outage probability can be
system can be obtained as
approximated as Eq. (7) shown on the top of the next page,
!2 where
N
X π
γmax = γ0 |h2,n ||h1,n | + |g| , (2) µ = p α1 α2 L 1 (−K1 )L 12 (−K2 ),
4 d1 d2 (K1 + 1)(K2 + 1) 2
n=1
and
where γ0 = P/N0 represents the transmit SNR.   2 
2
1 π L 12 (−K1 )L 12 (−K2 ) 
σ2 =

III. P ERFORMANCE A NALYSIS 1 − . (6)
dα1 α2
1 d2 16(K1 + 1)(K2 + 1)
In this section, we provide a detailed analysis of the achiev-
able systems performance. Specifically, two important metrics,
i.e., ergodic capacity and outage probability, are considered. Proof: See Appendix B.
We start with the ergodic capacity. Theorem 2 presents a closed-from approximation for the
outage probability, which consists of only elementary func-
tions, hence can be efficiently evaluated. Also, we observe
A. Ergodic Capacity that both µ and σ 2 are symmetric functions with respect to
The ergodic capacity of the system can be expressed as K1 and K2 , which implies K1 and K2 have identical impact
on the outage probability. In addition, although Theorem 2 is
C = E {log2 (1 + γmax )} . (3)
obtained with the assumption of large N , as will be shown
Since the exact distribution of γmax is intractable, it is chal- through numerical results, the approximation turns out to be
lenging to characterize the exact ergodic capacity. Therefore, sufficiently tight even for moderate N .
we resort to a tractable bound with the help of Jensen’s To gain further insight, we now look into the high SNR
inequality, and have the following key results: regime, and we have the following important result:
Theorem 1: The ergodic capacity of the system is upper Theorem 3: When γ0 → ∞, the outage probability can be
bounded by Cup given in Eq. (4), shown on the top of the next accurately approximated by
page, where L 12 (·) denotes the Laguerre polynomial [15]. "√ # −(N +1)
Proof: See Appendix A. πaN dα 1 N α2 N α3
1
d2 d3 2γ0
Pout,2 ≈ , (8)
Theorem 1 provides a closed-form expression involving Γ N + 32 (N + 1)! γth
only elementary functions, which is applicable for arbitrary where
system configurations, thereby enabling efficient evaluation 
of the ergodic capacity performance. In addition, it also a = (K1 + 1)(K2 + 1)e−(K1 +K2 ) Ei(K1 ) + Ei(K2 ) − 2γ
facilitates the characterization of the impact of key parameters s !
on the ergodic capacity. In particular, we have the following (K1 + 1)(K2 + 1)
observations: − ln K1 − ln K2 + 2K0 2 ,
dα1 α2
1 d2 γ0
Remark 1: The ergodic capacity upper bound increases
monotonically with N . For sufficiently large N , the ergodic γ denotes the Euler’s constant, Ei(x) is the exponential integral
capacity Cup is mainly dominated by the term log2 γ1 N 2 , [16], and Kn (x) is the modified Bessel function of the second
2
L 1 (−K1 )L 1 (−K2 )π kind [16].
where γ1 = γ0 16d2α1 dα2 (K 2+1)(K +1) , indicating an effective Proof: See Appendix C.
1 2 1 2
SNR gain of N 2 . This can be explained by the fact that IRS Theorem 3 suggests that a diversity order of N + 1 is
not only achieves the beamforming gain, but also attains the achieved. Moreover, the impact of LOS component is mainly
inherent aperture gain by collecting more signal power. reflected on the achievable coding gain of the system.
Remark 2: Observing that L 21 (−K) is a monotonically
increasing function of K, which suggests that a strong LOS IV. S IMULATION R ESULTS
component would enhance the ergodic capacity. In addition, In this section, numerical results are presented to validate
the ergodic capacity upper bound Cup is a symmetric function the theoretical analysis in the previous section. Unless oth-

1089-7798 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: UNIVERSITY OF BIRMINGHAM. Downloaded on July 25,2020 at 10:04:11 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/LCOMM.2020.3011843, IEEE
Communications Letters

( 
1 N
Cup = log2 1 + γ0 α3 + α1 α2 +
d3 d1 d2
s )
N (N − 1)π 2  2 N 2 π3
α1 α2 L 12 (−K1 )L 12 (−K2 ) + L 12 (−K1 )L 12 (−K2 ) (4)
16d1 d2 (K1 + 1)(K2 + 1) 8dα1 α2 α3
1 d2 d3 (K1 + 1)(K2 + 1)

qγ   q   q 
1 1  γ0 1
th
− Nµ −( γγth0 − N µ)2 γth
γ0 − N µ
Pout,1 ≈ + erf √ − p exp   1 + erf  p  .
2 2 2N σ 2 2 1 + N σ 2 dα
3
3 2(1/dα
3 + Nσ )
3 2
2N σ 2 (1 + N σ 2 dα
3 )
3

(7)

14 9
Upper bound, K=10
Monte Carlo, K=10
12 8
Upper bound, K=1
Monte Carlo, K=1
Ergodic capacity (bits/s/Hz)

Ergodic capacity (bits/s/Hz)

10 7
N=64

8 6 K = 100
N=256
K=1
6 5

N=32
4 4 N=64

2 3

0 2
−20 −15 −10 −5 0 5 10 60 80 100 120 140 160 180 200 220 240
Transmit power (dBm) d1 (m)

Fig. 2. Impact of K and N on the ergodic capacity. Fig. 3. Impact of IRS location on the ergodic capacity.

0
10

erwise specified, the following set of parameters are used in

the simulations. As in [5], we focus on the sub-6G scenario,
the system has a bandwidth of 180KHz, and the noise power −1
10
Outage probability

spectrum density is -173 dBm/Hz. The distances are set to be N=32

d1 = 150 m, d2 = 150 m and d3 = 200 m, with path loss N=64

exponents given by α0 = 3.5, α1 = α2 = 2.0, respectively.

The reference path loss at the reference distance d0 = 1 m is 10
−2

set to be −30 dB and the outage threshold γth = 10 dB. Also,

we set K1 = K2 = K. Approximation
Fig. 2 illustrates the ergodic capacity of the system with −3
Monte Carlo
10
different N . As can be readily observed, the ergodic capacity −30 −28 −26 −24
Transmit power P(dBm)
−22 −20 −18

upper bound is tight for all configurations, indicating the

accuracy of the closed-form expression in Theorem 1. In Fig. 4. Outage approximation Pout,1 with K = 1.
addition, we observe the intuitive result that the ergodic
capacity increases monotonically with the number of IRS
reflecting elements N and Rician factor K. decreasing rate is faster with a larger N .
Fig. 3 shows the impact of IRS location on the ergodic Fig. 5 validates the accuracy of Pout,2 in the high SNR
capacity assuming the IRS is deployed between the line regime when K = 1. As readily can be seen, the high SNR
segment of T and R, i.e., d1 + d2 = 300 m. As can be readily approximation is quite accurate, thus validates the correctness
observed, the ergodic capacity is a symmetric function with of Pout,2 . Also, a diversity order of N + 1 is observed,
respect to d1 and d2 . The minimum is achieved when the IRS indicating the great benefit of increasing N in terms of outage
is deployed in the middle of the T and R. The above result performance.
indicates that it better to deploy the IRS in the vicinity of
either T or R to obtain higher capacity.
V. CONCLUSION
Fig. 4 plots the outage probability of the system with K =
1. It can be observed that the approximation works well, even This paper has studied the ergodic capacity and outage prob-
for moderate value of N . In addition, we see that the outage ability of IRS-aided SISO systems under mixed Rayleigh and
probability decreases sharply when the SNR increases, and the Rician fading channels. Closed-from expressions are derived

1089-7798 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: UNIVERSITY OF BIRMINGHAM. Downloaded on July 25,2020 at 10:04:11 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/LCOMM.2020.3011843, IEEE
Communications Letters

 
PN PN
10
0

Monte Carlo
Regarding E n=1 j=1 |h2,n ||h1,n ||h2,j ||h1,j | , noticing
Approximation j6=n
that, for the Rician variable |hl,n |, l ∈ {0, 1},
−1 r
10
π
E{|hl,n |} = αl L 1 (−Kl ), (12)
Outage probability

4dl (Kl + 1) 2
−2
N=1
10
and observing the fact that h1 and h2 are independent, we
obtain
 
−3
10 N=2  

X N X N 

E |h2,n ||h1,n ||h2,j ||h1,j |

 

10
−4

−10 −5 0 5 10 15 20
n=1 j=1 
Transmit power P(dBm)
j6=n
N (N − 1)π 2  2  2
= L 1 (−K1 ) L 1 (−K2 ) .
Fig. 5. High SNR outage approximation Pout,2 with K = 1. 16dα1 α2
1 d2 (K1 + 1)(K2 + 1)
2 2

(13)
3) Computing x3 : For the Rayleigh variable |g|, we have
for the ergodic capacity upper bound and outage approxima- r
tion, which provide efficient means to evaluate the system π
E{|g|} = . (14)
performance. Moreover, concise expressions are obtained in 2dα3
3

the asymptotic regime, which sheds lights on the impact of

As such, x3 can be calculated as
key parameters on the system performance. It is revealed that s
the use of IRS contributes to an effective SNR gain of N 2 , and π3 N 2
the diversity order can be increased to N +1. Also, the position x3 = α1 α2 α3 L 1 (−K1 )L 12 (−K2 ).
8d1 d2 d3 (K1 + 1)(K2 + 1) 2
of IRS also has significant impact on the system performance,
(15)
it is desirable to deploy the IRS close to the transmitter or
receiver, and with strong LOS path. To this end, pulling x1 , x2 and x3 together yields the desired
result.
A PPENDIX A
P ROOF OF T HEOREM 1 A PPENDIX B
Applying the Jensen’s inequality, we have the upper bound P ROOF OF T HEOREM 2
of the capacity as According to the definition, the outage probability can be
transformed into
C ≤ Cup = log2 (1 + E {γmax }) . (9)  r 
γth
Pout (γth ) = Prob z ≤ , (16)
The remaining task is to compute E {γmax }. Using the rela- γ0
tionship in (2) and applying the binomial expansion theorem, PN
we have where z , u+|g| with u , n=1 |h2,n ||h1,n |. For sufficiently
 !2  large N , u can be accurately approximated by the normal
 X N 
 2 distribution according to the central limit theorem, i.e., u ∼
E {γmax } = E |g| + E |h2,n ||h1,n | +
| {z }   N (N µ, N σ 2 ), where
n=1
x1 | {z } µ = E{|h2,n ||h1,n |} (17)
x2
( ) π
N
X = p α1 α2 L 1 (−K1 )L 12 (−K2 ),
2E |h2,n ||h1,n ||g| . (10) 4 d1 d2 (K1 + 1)(K2 + 1) 2
n=1 and
| {z }
x3 1
σ 2 = D{|h2,n ||h1,n |} = α1 α2 (18)
We now calculate x1 , x2 and x3 one-by-one. d1 d2
1) Computing x1 : Obviously, we have x1 = dα13 .
 
π2  2
3
2) Computing x2 : Using the binomial expansion, x2 can be 1− L 12 (−K1 )L 12 (−K2 ) .
16(K1 + 1)(K2 + 1)
further expanded as
(N ) Since z is the sum of independent random variable u and |g|,
X its cumulative distribution function (CDF) can be calculated
x2 =E |h2,n |2 |h1,n |2 + via
n=1
Z ∞


 
 Fz (z) ≈ fu (z − x)F|g| (x)dx, (19)
XN XN  −∞
E |h2,n ||h1,n ||h2,j ||h1,j | . (11) h i
−(x−N µ)2

 
 where fu (x) = √ 1 exp is the probability
n=1 j=1  2πN σ2 2N σ2  α3 2 
j6=n
−d3 x
nP o density function (PDF) of u and F|g| (x) = 1 − exp 2
N 2 2 N
It is easy to see that E n=1 |h2,n | |h1,n | = α α
d1 1 d2 2
. denotes the CDF of Rayleigh variable |g|.

1089-7798 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: UNIVERSITY OF BIRMINGHAM. Downloaded on July 25,2020 at 10:04:11 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/LCOMM.2020.3011843, IEEE
Communications Letters

To this end, the desired result can be obtained after some 3) For n = p, we observe that Pn=p is dominated by the
algebraic manipulations. term n = p = 0, thus we have
Pn=p ≈2(K1 + 1)(K2 + 1)e−(K1 +K2 ) ×
A PPENDIX C  p 
P ROOF OF T HEOREM 3 K0 2 (K1 + 1)(K2 + 1)β . (27)
Recall γmax in Eq. (2), it can be interpreted as the effec- Combining the above three parts, we have t = 0 and
tive SNR of an equal gain combining SIMO system. Then, 
according to [17], in order to obtain the outage approximation a =(K1 + 1)(K2 + 1)e −(K1 +K2 )
Ei(K1 ) + Ei(K2 ) − 2γ
in the high SNR regime, it is sufficient to characterize the
s !
behavior of the probability density function (PDF) of the SNR
(K1 + 1)(K2 + 1)
of individual branch near the origin. Specifically, let β denote − ln K1 − ln K2 + 2K0 2 .
the SNR of the branch, f (β) denote the PDF of β. If dα 1 α2
1 d2 γ0
(28)
lim fβ (β) = aβ t + O(β t+ε ), (20)
β→0 To this end, invoking [17, Eq. (12)], the desired results can
where O(x) is the big O notation, then the high SNR outage be obtained after some algebraic manipulations.
probability can be accurate characterized by the parameters a
and t. R EFERENCES
Therefore, the main task is to obtain the behavior of the PDF [1] C. Huang, et al, “Holographic MIMO surfaces for 6G wireless net-
works: Opportunities, challenges, and trends”, [Online]. Available: http-
of |g|2 and |h2,n |2 |h1,n |2 near the origin. Since |g|2 follows s://arxiv.org/abs/1911.12296.
the exponential distribution, its behavior near the origin can [2] C. Huang, R. Mo, and C. Yuen, “Reconfigurable intelligent surface as-
be easily obtained. Hence, we focus on |h2,n |2 |h1,n |2 in the sisted multiuser MISO systems exploiting deep reinforcement learning”,
[Online]. Available:https://arxiv.org/abs/2002.10072.
following. [3] Q. Wu and R. Zhang, “Intelligent reflecting surface enhanced wireless
Capitalizing on the results of [18], the PDF of random network via joint active and passive beamforming,” IEEE Trans. Wireless
variable β = dα 1 α2 2 2
1 d2 |h2 | |h1 | can be derived as
Commun., vol. 18, no. 11, pp. 5394C-5409, Nov. 2019.
[4] Q. Wu and R. Zhang, “Beamforming optimization for wireless network
n p !2 aided by intelligent reflecting surface with discrete phase shifts,” accept-
∞
−(K1 +K2 )
X K12 K22 ed to appear in IEEE Trans. Commun., 2019.
fβ (β) = 2(K1 + 1)(K2 + 1)e × [5] C. Huang, et al, “Reconfigurable intelligent surfaces for energy efficien-
n,p=0
n!p!
cy in wireless communication,” IEEE Trans. Wireless Commun., vol. 18,
p n+p  p  no. 8, pp. 4157-C4170, Aug. 2019.
(K1 + 1)(K2 + 1)β Kn−p 2 (K1 + 1)(K2 + 1)β . [6] Y. Han, W. Tang, S. Jin, C. Wen, and X. Ma, “Large intelligent surface-
sssisted wireless communication exploiting statistical CSI,” IEEE Trans.
With the following relationships of bessel-K function [16, Eq. Veh. Technol., vol. 68, no. 8, pp. 8238–8242, Aug. 2019.
[7] S. Zhang and R. Zhang, “Capacity characterization for intelligent reflect-
10.30.2] ing surface aided MIMO communication,” IEEE J. Sel. Areas Commun.,
1 1 to appear. [Online]. Available: https://arxiv.org/abs/1910.01573.
lim Kν (z) ≈ Γ(ν)( z)−ν , ℜ ν > 0, (21) [8] X. Hu, C. Zhong, Y. Zhu, X. Chen, and Z. Zhang, “Programmable
z→0 2 2 metasurface based multicast systems: Design and analysis,” accepted
Kν (z) = K−ν (z), (22) to appear in IEEE J. Selected Areas Commun., 2020.
[9] G. Zhou, C. Pan, H. Ren, K. Wang, and A. Nallanathan, “Intelligent
where ℜ ν represents the real part of ν. We find it is convenient reflecting surface sided multigroup multicast MISO communication
systems,” [Online]. Available: https://arxiv.org/abs/1909.04606.
to consider three separate cases depending on the relationship [10] Y. Jia, C. Ye, and Y. Cui, “Analysis and optimization of an intelligent
of n and p, namely, reflecting surface-assisted system with interference,” [Online]. Available:
https://arxiv.org/pdf/2002.00168.
lim fβ (β) = Pn>p + Pn<p + Pn=p . (23) [11] C. Guo, Y. Cui, F. Yang, and L. Ding, “Outage probability analysis and
β→0 minimization in intelligent reflecting surface-assisted MISO systems,”
IEEE Commun. Lett., Early Access.
1) For n > p, we have [12] D. Kudathanthirige, D. Gunasinghe, and G. Amarasuriya, “Performance
∞ X
X ∞ analysis of intelligent reflective surfaces for wireless communication,”
[Online]. Available: https://arxiv.org/abs/2002.05603.
Pn>p ≈ 2(K1 + 1)(K2 + 1)e−(K1 +K2 ) (24) [13] I. Trigui, W. Ajib, and W. Zhu, “A comprehensive study of reconfig-
p=0 n=p+1 urable intelligent surfaces in generalized fading,” [Online]. Available:
n p !2 https://arxiv.org/abs/2004.02922.
K12 K22 p 2p 1 [14] J. Gao, C. Zhong, X. Chen, H. Lin, and Z. Zhang, “Unsupervised
(K1 + 1)(K2 + 1)β Γ(n − p). learning for passive beamforming,” IEEE Wireless Commun. Lett., vol.
n!p! 2 24, no. 5, pp. 1052-1056, May 2020.
[15] M. Abramowitz, and I. A. Stegun, Handbook of mathematical functions.
It is easy to see that Pn>p is mainly determined by p = 0, New York: Dover Publication Inc., 1974.
hence we have [16] F. W. Olver, D. W. Lozier, R. F. Boisvert, and C. W. Clark, NIST
∞ Handbook of Mathematical Functions. New York: Cambridge University
X K1n Press, 2010.
Pn>p ≈ 2(K1 + 1)(K2 + 1)e−(K1 +K2 ) . (25) [17] Z. Wang and G. B. Giannakis, “A simple and general parameterization
n=1
2nn! quantifying performance in fading channels,” IEEE Trans. Commun.,
vol. 51, no. 8, pp. 1389–1398, Aug. 2003.
2) For n < p, exploiting the symmetric property in Eq. (22), [18] N. O’Donoughue and J. M. F. Moura, “On the product of independent
we can similarly derive complex Gaussians,” IEEE Trans. Signal Process., vol. 60, no. 3, pp.
∞
1050–1063, March 2012.
X K2p
Pn<p ≈ 2(K1 + 1)(K2 + 1)e−(K1 +K2 ) . (26)
p=1
2pp!

1089-7798 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: UNIVERSITY OF BIRMINGHAM. Downloaded on July 25,2020 at 10:04:11 UTC from IEEE Xplore. Restrictions apply.