Di Renzo and Song EURASIP Journal on Wireless Communications and

Networking (2019) 2019:99

https://doi.org/10.1186/s13638-019-1403-7

RESEARCH Open Access

Reflection probability in wireless

networks with metasurface-coated
environmental objects: an approach based
on random spatial processes
Marco Di Renzo* and Jian Song

Abstract
An emerging and promising vision of wireless networks consists of coating the environmental objects with
reconfigurable metasurfaces that are capable of modifying the radio waves impinging upon them according to the
generalized law of reflection. By relying on tools from point processes, stochastic geometry, and random spatial
processes, we model the environmental objects with a modified random line process of fixed length and with random
orientations and locations. Based on the proposed modeling approach, we develop the first analytical framework that
provides one with the probability that a randomly distributed object that is coated with a reconfigurable metasurface
acts as a reflector for a given pair of transmitter and receiver. In contrast to the conventional network setup where the
environmental objects are not coated with reconfigurable metasurfaces, we prove that the probability that the typical
random object acts as a reflector is independent of the length of the object itself. The proposed analytical approach is
validated against Monte Carlo simulations, and numerical illustrations are given and discussed.
Keywords: Wireless networks, Reconfigurable metasurfaces, Stochastic geometry, Random spatial processes,
Reflection probability

1 Methods/experimental enable its time critical and energy-efficient delivery,

The methods used in the paper are based on the mathe- and of accurately localizing people and objects in harsh
matical tools of random spatial processes and stochastic propagation environments. Future wireless networks will
geometry. A new analytical framework for performance have to fulfill the challenging requirement of intercon-
analysis is introduced. The theoretical framework is vali- necting the physical and digital worlds in a seamless and
dated against Monte Carlo simulations. sustainable manner [1, 2].
To fulfill these challenging requirements, it is appar-
2 Introduction ent that it is not sufficient anymore to rely solely on
Future wireless networks will be more than allowing people, wireless networks whose logical operation is software-
mobile devices, and objects to communicate (https://www. controlled and optimized [3]. The wireless environment
comsoc.org/publications/ctn/what-will-6g-be). Future itself needs to be turned into a software-reconfigurable
wireless networks will be turned into a distributed intel- entity [4], whose operation is optimized to enable unin-
ligent wireless communications, sensing, and computing terrupted connectivity. Future wireless networks need a
platform, which, besides communications, will be capable smart radio environment, i.e., a wireless environment that
of sensing the environment to provide context-awareness is turned into a reconfigurable space that plays an active
capabilities, of locally storing and processing data to role in transferring and processing information.
Different solutions towards realizing this wireless future
*Correspondence: [email protected] are currently emerging [5–13]. Among them, the use of
Laboratoire des Signaux et Systèmes, CNRS, CentraleSupelec, Univ Paris-Sud,
Université Paris-Saclay, Plateau du Moulon, 91192 Gif-sur-Yvette, France
reconfigurable metasurfaces constitutes a promising and
enabling solution to fulfill the challenging requirements

© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the
Creative Commons license, and indicate if changes were made.
Di Renzo and Song EURASIP Journal on Wireless Communications and Networking (2019) 2019:99 Page 2 of 15

of future wireless networks [14]. Metasurfaces are thin objects can optimize the reflected signals in anomalous
metamaterial layers that are capable of modifying the directions beyond Snell’s law. The corresponding achiev-
propagation of the radio waves in fully customizable ways able performance and the associated optimal setups are
[15], thus owing the potential of making the transfer and unknown.
processing of information more reliable [16]. Also, they Motivated by these considerations, we develop an ana-
constitute a suitable distributed platform to perform low- lytical framework that allows one to quantify the probabil-
energy and low-complexity sensing [17], storage [18], and ity that a random object coated with reconfigurable meta-
analog computing [19]. For this reason, they are partic- surfaces acts as a reflector for a given pair of transmitter
ularly useful for improving the performance of non-line- and receiver. Even though reconfigurable metasurfaces
of-sight transmission, e.g., to appropriately customize the can be used to control and customize the refractions from
impact of multipath propagation. environmental objects, in the present paper, we focus our
In [13], in particular, the authors have put forth a net- attention on controlling and customizing only the reflec-
work scenario where every environmental object is coated tions of signals, since refractions may be subject to severe
with reconfigurable metasurfaces, whose response to the signal’s attenuation. Our proposed approach, in particu-
radio waves is programmed in software by capitalizing on lar, is based on modeling the environmental objects with
the enabling technology and hardware platform currently a modified random line process of fixed length and with
being developed in [20]. Current research efforts towards random orientations and locations. In contrast to the con-
realizing this vision are, however, limited to implement ventional network setup where the environmental objects
hardware testbeds, e.g., reflect-arrays and metasurfaces, are not coated with reconfigurable metasurfaces, we prove
and on realizing point-to-point experimental tests [5–13]. that the probability that the typical random object acts as
To the best of the authors’ knowledge, notably, there exists a reflector is independent of the length of the object itself.
no analytical framework that investigates the performance The proposed analytical approach is validated against
of large-scale wireless networks in the presence of recon- Monte Carlo simulations, and numerical illustrations are
figurable metasurfaces. In the present paper, motivated given and discussed. In the present paper, we limit our-
by these considerations, we develop the first analytical selves to analyze 2D network scenarios, but our approach
approach that allows one to study the probability that a can be applied to 3D network topologies as well. This non-
random object coated with a reconfigurable metasurface trivial generalization is postponed to a future research
acts as a reflector according to the generalized laws of work.
reflection [15]. To this end, we capitalize on the mathe- The remainder of the present paper is organized as
matical tool of random spatial processes [21, 22]. follows. In Section 3, the system model is introduced.
Random spatial processes are considered to be the most In Section 4, the problem is formulated in mathemati-
suitable analytical tool to shed light on the ultimate per- cal terms. In Section 5, the analytical framework of the
formance limits of innovative technologies when applied reflection probability is described. In Section 6, numerical
in wireless networks and to guide the design of opti- results are illustrated, and the proposed approach is vali-
mal algorithms and protocols for attaining such ultimate dated against Monte Carlo simulations. Finally, Section 7
limits. Several recent results on the application of ran- concludes the paper.
dom spatial processes in wireless networks can be found Notation: The main symbols and functions used in this
in [23–31]. Despite the many results available, however, paper are reported in Table 1.
fundamental issues remain open [28]. In the current lit-
erature, in particular, the environmental objects are mod-
Table 1 Main symbols and functions used throughout the paper
eled as entities that can only attenuate the signals, by
Symbol/function Definition
making the links either line-of-sight or non-line-of-sight,
e.g., [25–27]. Modeling anything else is acknowledged Pr{A} Probability of Event A
 
to be difficult. Just in [32], the authors have recently Pr A Probability of complement of Event A
investigated the impact of reflections, but only based H (·), H̄ (·) Heaviside function, complementary Heavi-
on conventional Snell’s laws. This work highlights the side function
analytical complexity, the relevance, and the non-trivial (xTx , yTx ) Location of the transmitter
performance trade-offs: The authors emphasize that the (xRx , yRx ) Location of the receiver
obtained trends highly depend on the fact that the total  
xobject , yobject Location of the center of the typical object
distance of the reflected paths is almost always two times
larger than the distance of the direct paths. This occurs (xend1 , yend1 ), (xend2 , yend2 ) Coordinates of the end points of the typical
object
because the angles of incidence and reflection are the
L Length of the typical object
same based on Snell’s law. In the presence of recon-
figurable metasurfaces, on the other hand, the random Rnet Radius of the network
Di Renzo and Song EURASIP Journal on Wireless Communications and Networking (2019) 2019:99 Page 3 of 15

3 System model • Scenario II: The second scenario corresponds to the

We consider a wireless network on a bi-dimensional plane, case study where the typical object is not coated with
where the transmitters and receivers are distributed inde- a reconfigurable metasurface. This is the
pendently of each other. Without loss of generality, the state-of-the-art scenario, where the angle of
location of the transmitter and receiver of interest, i.e., the reflection needs to be equal to the angle of incidence
probe transmitter and receiver, are denoted by (xTx , yTx ) according to Snell’s law of reflection [15].
and (xRx , yRx ), respectively.
Besides the transmitter and receiver, we assume that The aim of the present paper is to develop a mathe-
environmental objects, e.g., buildings in a urban outdoor matical theory to compute the probability that the typical
scenario, are randomly distributed in the same region. An object can act as a reflector, i.e., the reflection probability,
example of the network model is depicted in Fig. 1. More and to quantify the gain of adding reconfigurable metasur-
precisely, the environmental objects are assumed to fol- faces in wireless networks. For analytical tractability, we
low a Boolean model of line segments with the following assume that the reconfigurable metasurfaces are capable
properties [21]: of producing any angle of reflection for any given loca-
tion of transmitter and receiver, for any angle of incidence,
• The center points of the objects form a homogeneous and for any length. This yields the best-case performance
Poisson point process. bound compared with conventional Snell’s law of reflec-
• The orientation of the objects are independent and tion. The analysis, in addition, is conducted by relying on
identically distributed in [0, 2π]. ray tracing arguments, in order to highlight the poten-
• The lengths of the objects are fixed and all equal to L. tial gains of using reconfigurable metasurfaces. The two
• The random orientation and the center points of the case studies are sketched in Fig. 2. Generalizations of the
objects are independent of each other. proposed analytical framework are left to future research
works.
We consider a generic environmental  object, i.e.,  the
typical object, and denote its center by xobject , yobject , and 4.1 Scenario I: Reflections in the presence of
the coordinates of its two end points by (xend1 , yend1 ) and reconfigurable metasurfaces
(xend2 , yend2 ). In the presence of reconfigurable metasurfaces, an arbi-
trary angle of reflection can be obtained for any angle
4 Problem formulation of incidence. This implies that the typical object acts
The objective of the present paper is to compute the prob- as a reflector for a transmitter and receiver if they are
ability that a randomly distributed object can act as a both located on the same side of the infinite line passing
reflector for the pair of transmitter and receiver located in through the end points (xend1 , yend1 ) and (xend2 , yend2 ) of
(xTx , yTx ) and (xRx , yRx ), respectively. We analyze two case the typical object.
studies: For ease of exposition, we introduce the following event.
• Scenario I: The first scenario corresponds to the case
study where the typical object is coated with a Event 1 The probe transmitter, Tx, and receiver, Rx, are
reconfigurable metasurface, which can optimize the located on the same side of the infinite line passing through
angle of reflection regardless of the angle of incidence the end points (xend1 , yend1 ) and (xend2 , yend2 ) of the typical
[15]. object.

Fig. 1 Probe transmitter (Tx) and receiver (Rx) in the presence of randomly distributed environmental objects
Di Renzo and Song EURASIP Journal on Wireless Communications and Networking (2019) 2019:99 Page 4 of 15

Fig. 2 Scenario I: a In the presence of reconfigurable metasurfaces, the angle of incidence and reflection are not necessarily the same. Scenario II: b
According to Snell’s law of reflection, the angle of incidence and reflection are the same

Therefore, the typical object acts as a reflector if Event 1 4.1.2 Approach 2

holds true. Our objective is to formulate the probability The probability that Event 1 holds true can be formu-
of Event 1, i.e., to compute Pr {Event 1}. This latter prob- lated also in terms of the positions of the transmitter and
ability can be formulated in two different but equivalent receiver with respect to the infinite line that connects the
ways. end points of the typical object. In particular, Event 1
holds true if both the transmitter and receiver are on the
4.1.1 Approach 1 same side of the infinite line, i.e., either they are above or
Let us consider the infinite line that connects the loca- they are below the infinite line. This interpretation can be
tions of transmitter and receiver, and the infinite line that viewed as a problem of classifying points with respect to a
connects the two end points of the typical object. Event 1 line. This interpretation is depicted in Fig. 4.
holds true if the intersection point, denoted by (x∗ , y∗ ), of In mathematical terms, therefore, Pr {Event 1} can be
these two infinite lines falls outside the line segment that formulated as follows:
connects that transmitter and the receiver. An illustration
is given in Fig. 3.
In mathematical terms, therefore, Pr {Event 1} can be  
{[Tx is above the line] ∩ [Rx is above the line]}
formulated as follows: Pr {Event 1} = Pr
∪ {[Tx is below the line] ∩ [Rx is below the line]}
 
Pr {Event 1} = 1 − Pr Event 1 (1) (3)

where Event 4.2 Scenario II: Reflections in the absence of

  1 is the complement of Event 1, and reconfigurable metasurfaces
Pr Event 1 denotes the probability that the intersection
In the absence of reconfigurable metasurfaces, the typical
point (x∗ , y∗ ) is on the Tx-Rx line segment: object acts as a reflector, for a given transmitter and
    receiver, only if the angles of reflection and incidence are
min (xTx , xRx )  x∗  max (xTx , xRx ) the same. This is agreement with Snell’s law of reflection,
Pr Event 1 = Pr ∗
∩ min (yTx , yRx )  y  max (yTx , yRx ) and imposes some geometric constraints among the
(2) locations of the typical object, the transmitter, and the

a b
Fig. 3 Illustration of Event 1 based on Approach 4.1.1. a Event 1 holds true if the intersection point between the infinite lines falls outside the Tx-Rx
line segment. b Otherwise, Event 1 does not hold true
Di Renzo and Song EURASIP Journal on Wireless Communications and Networking (2019) 2019:99 Page 5 of 15

a b c
Fig. 4 Illustration of Event 1 based on Approach 4.1.2. Event 1 holds true if both Tx and Rx are above a or below b the infinite line corresponding to
the typical object. Event 1 does not hold true if Tx and Rx are not on the same side of the line c

receiver. In order to compute the corresponding probabil- Based on Snell’s law of reflection, therefore, the typical
ity of occurrence, we introduce the following event. object acts a reflector if the following event holds true.

Event 3 The transmitter and receiver are located on the

Event 2 The mid-perpendicular of the line segment that same side of the infinite line passing through the end points
connects the transmitter and receiver intersects the line of the typical object, and the mid-perpendicular of the line
segment that represents the typical object. segment that connects the transmitter and receiver inter-
sects the line segment that represents the typical object.
Let (x∗ , y∗ ) denote the intersection between the mid-
perpendicular of the line segment that connects the An illustration is given in Fig. 5. In mathematical terms,
transmitter and receiver, and the line segment that rep- the probability of occurrence of Event 3 can be formulated
resents the typical object. According to Snell’s law of as follows:
reflection, for some given locations of the transmitter
and receiver, the typical object acts as a reflector if the Pr {Event 3} = Pr {Event 1 ∩ Event 2} (5)
mid-perpendicular of the line segment that connects the
transmitter and receiver intersects the line segment that 5 Analytical formulation of the reflection
represents the typical object (i.e., Event 2), and, at the probability
same time, the transmitted and receiver are located on In this section, we introduce analytical expressions of the
the same side of the infinite line passing through the end probability of occurrence of the three events introduced
points of the typical object (i.e., Event 1). in the previous sections, and, therefore, characterize the
In mathematical terms, the probability of occurrence of probability that the typical object acts as a reflector in the
Event 2 can be formulated as follows: presence and in the absence of reconfigurable metasur-
faces. First, we begin with some preliminary results.
 
min (xend1 , xend2 )  x∗  max (xend1 , xend2 ) 5.1 Preliminary results
Pr {Event 2} = Pr
∩ min (yend1 , yend2 )  y∗  max (yend1 , yend2 ) Lemma 1 Let (xTx , yTx ) and (xRx , yRx ) be the locations
(4) of the probe transmitter and receiver, respectively. The

Fig. 5 Illustration of Event 3. In a, Event 1 and Event 2 hold true: The typical object acts as a reflector. In b, Event 1 holds true but Event 2 does not
hold true: The typical object cannot be a reflector
Di Renzo and Song EURASIP Journal on Wireless Communications and Networking (2019) 2019:99 Page 6 of 15

infinite line passing through them can be formulated as a Poisson point process with random orientations, which
√
follows: implies p = Rnet u and α ∈ [0, 2π]. The rest follows from
geometric considerations.
y = mx + z (6)
yTx −yRx
where m = xTx −xRx , and z = yRx − mxRx . Lemma 3 The mid-perpendicular of the infinite line in
( 6) is as follows:
Proof It follows by definition of line passing through two
points. y = mp x + zp (9)
1 1
Lemma 2 Let us consider the typical object of length L where mp = − m , and zp = 2m (xTx + xRx )+ 12 (yTx + yRx ).
depicted in Fig. 6. The distance between the center of the
√ Proof See Appendix 1.
line segment and the origin is p = Rnet u, where u is a
uniformly distributed random variable in [0, 1], and Rnet
is largest size of the region of interest. Let α be the angle Lemma 4 The intersection point between the infinite line
between the perpendicular line to the line segment, which that connects the transmitter and the receiver, and the
passes through the center of the object, and the horizontal infinite line that connects the end points of the line seg-
axis. The infinite line passing through the end points of the ment representing the typical object can be formulated as
object can be formulated as follows: follows:

x cos α + y sin α = p (7) p − z sin α

x∗ =
m sin α + cos α (10)
where α ∈ [0, 2π].
y∗ = mx∗ + z
In addition,
 the center
 of the line segment can be writ-
ten as xobject , yobject = (p cos α, p sin α), and its end The intersection point between the (infinite) mid-
points (xend1 , yend1 ) and (xend2 , yend2 ) can be formulated perpendicular line to the line segment that connects the
as follows: transmitter and the receiver, and the infinite line that con-
L L nects the end points of the line segment representing the
xend1 = xobject − sin α, yend1 = yobject + cos α typical object can be formulated as follows:
2 2
L L
xend2 = xobject + sin α, yend2 = yobject − cos α p − zp sin α
2 2 x∗ =
mp sin α + cos α (11)
(8)
y∗ = mp x∗ + zp
Proof The proof follows by noting that the centers of the
line segments (the objects) are distributed according to Proof Equation 10 follows by solving the system of
equations in (6) and (7). Equation 11 follows by solving the
system of equations in (7) and (9).

Lemma 5 Let a generic infinite line formulated as:

ax + by + c = 0. The following holds true:

• The point (x1 , y1 ) is above the line if ax1 + by1 + c > 0

and b > 0, or if ax1 + by1 + c < 0 and b < 0.
• The point (x1 , y1 ) is below the line if ax1 + by1 + c < 0
and b > 0, or if ax1 + by1 + c > 0 and b < 0.

Proof See Appendix 2.

5.2 Scenario I: Reflection probability in the presence of

reconfigurable metasurfaces
Theorems 1 and 2 provide one with analytical expres-
sions of the probability that the typical object acts as
a reflector if it is coated with reconfigurable metasur-
Fig. 6 Geometric representation of the typical object in polar faces. Theorem 1 is computed based on Approach 1, and
coordinates
Theorem 2 based on the Approach 2.
Di Renzo and Song EURASIP Journal on Wireless Communications and Networking (2019) 2019:99 Page 7 of 15

Theorem 1 Based on Approach 1, the probability of and

occurrence of Event 1 is as follows:
ρ2 (α, xTx , yTx , xRx , yRx )
!   2 "
  xTx cos α + yTx sin α xRx cos α + yRx sin α
= 1 − max , ,0
Pr {Event1} = 1 − Pr Event1 Rnet Rnet
⎧ δ1  ⎫   
   x cos α + yTx sin α xRx cos α + yRx sin α
1 ⎨ 0 δ11  Tx Rx Tx Rx  δ2 2π1  Tx Rx Tx Rx  ⎬
θ α, x , x , y , y dα + 2π θ α, x , x , y , y dα × H 1 − max Tx , ,0
Rnet Rnet
= 1− + θ2 α, xTx , xRx , yTx , yRx dα + δ θ2 α, xTx , xRx , yTx , yRx dα
2π ⎩ 0δ2  
2
δ2   ⎭ (15)
+ δ θ3 α, xTx , xRx , yTx , yRx dα + + δ θ4 α, xTx , xRx , yTx , yRx dα
1 1
(12) Proof See Appendix 4.

where the integral

 limits are defined as Remark 1 Theorems 1 and 2 are two analytical formu-
√
δ1 = 2tan−1 m + 1 + m2 , and δ2 = lations of the same event. In the sequel, we show that they
 √  coincide. 
2π + 2tan−1 m − 1 + m2 , and the auxiliary functions
are given in Table 2. 5.3 Scenario II: Reflection probability in the absence of
reconfigurable metasurfaces
Proof See Appendix 3. The probability of occurrence of Event 3 is not easy to
compute. The reason is that Event 3 is formulated in terms
Theorem 2 Based on Approach 2, the probability of of the intersection of Events 1 and 2, which are not inde-
occurrence of Event 1 is as follows: pendent. In order to avoid the analytical complexity that
originates from the correlation between Events 1 and 2,
 2π
1 we propose a upper-bound to compute the probability
Pr {Event1} = ρ1 (α, xTx , yTx , xRx , yRx )dα of occurrence of Event 3. Before stating the main result,
2π 0
 2π
(13) we introduce the following proposition that provides one
1
+ ρ2 (α, xTx , yTx , xRx , yRx )dα with the probability of occurrence of Event 2.
2π 0

Proposition 1 The probability of occurrence of Event 2

where the integrand functions are defined as: can be formulated as follows.
# $
ρ1 (α, xTx , yTx , xRx , yRx ) min (xend1 , xend2 )  x∗  max (xend1 , xend2 )
Pr {Event2} = Pr
  2 ∩ min (yend1 , yend2 )  y∗  max (yend1 , yend2 )
x cos α + yTx sin α xRx cos α + yRx sin α # $
= min Tx , ,1 (14) 2π  3π  π  π
Rnet Rnet 1 2 2
   = 1 (α)dα + 2 (α)dα + 3 (α)dα + 4 (α)dα
xTx cos α + yTx sin α xRx cos α + yRx sin α 2π 3π
π 0 π
× H min , ,1 2 2

Rnet Rnet (16)

where 1 (α) = 1 (α) + 1 (α) + 1 (α) + 1 (α),
a b c d

Table 2 Auxiliary functions used in Theorem 1 2 (α) = 2 (α) +

a
2 (α) +
b
2 (α) +
c
2 (α),
d

3 (α) = 3 (α) + 3 (α) + 3 (α) + 3 (α), and

a b c d
Function definition
1 ([m sin α + cos α] ξ + z sin α) 4 (α) = 4
a
(α) + b
4 (α) + c
4 (α) + d
4 (α), which are all
f (α, ξ ) =
Rnet
  defined in Table 3.
g (α, ω) = 1 [m sin α+cos α][ω−z] + z sin α
⎛ Rnet
⎞ m


⎜  μ1 μ2 ⎟           
⎜α  ⎟= min μ , μ , 1 2 − max μ , μ , 0 2 H min μ , μ , 1 −max μ , μ , 0
⎝ ⎠ 1 2 3 4 1 2 3 4
μ μ
 3 4 Proof See Appendix 5.
⎛  ⎞
      

  ⎜  f α, max xTx , xRx g α, max yTx , yRx ⎟
θ1 α, xTx , xRx , yTx , yRx = ⎝ ⎜α  ⎟ × H (m)
 
 f α, min x , x  g α, min y , y 
⎠ Theorem 3 The probability of occurrence of Event 3 is
 Tx Rx Tx Rx ⎞
⎛ 
      
upper-bounded as follows:

  ⎜  f α, max xTx , xRx g α, min yTx , yRx ⎟
θ2 α, xTx , xRx , yTx , yRx = ⎜  ⎟ × H̄ (m)
α
⎝  
 f α, min x , x  g α, max y , y 
⎠ Pr {Event 3}  min {Pr {Event 1} , Pr {Event 2}} (17)
 Tx Rx Tx Rx ⎞
⎛ 
 
 
 
⎜  f α, min xTx , xRx
   
g α, min yTx , yRx ⎟ where Pr {Event 1} is formulated in Theorem 1 or
θ3 α, xTx , xRx , yTx , yRx = ⎜ 
⎝α  
⎟ × H (m)

 f α, max x , x

  
g α, max yTx , yRx
 ⎠ Theorem 2, and Pr {Event 2} is given in Proposition 1.
⎛  Tx Rx ⎞
      

  ⎜  f α, min xTx , xRx g α, max yTx , yRx ⎟
θ4 α, xTx , xRx , yTx , yRx = ⎜ α 
⎝  
⎟ × H̄ (m)
⎠ Proof : The proof follows by applying the Frechet
 f α, max x , x  g α, min y , y 
 Tx Rx Tx Rx inequality [33].
Di Renzo and Song EURASIP Journal on Wireless Communications and Networking (2019) 2019:99 Page 8 of 15

Table 3 Auxiliary functions used in Proposition 1

Function definition
  −1
1 z sin α 1
F (α, t) = Rnet t + mp sinp α+cos α mp sin α+cos α − cos α
  −1
1 mp zp sin α mp
G (α, v) = Rnet v + mp sin α+cos α − zp mp sin α+cos α − sin α
⎛      ⎞
    
 F α, − 2L sin α G α, 2L cos α
⎝ 
α     ⎠ H m sin α+cos
1 mp
1 (α) = − cos α H mp sin α+cos α − sin α
a
α
 F α, 2 sin α G α, − 2 cos α
L L p

⎛     ⎞
    
 F α, − L sin α G α, − 2L cos α
⎝α   2    ⎠ H m sin α+cos
1 mp
1 (α) = − cos α H̄ mp sin α+cos α − sin α
b
 α
 F α, 2 sin α
L L
G α, 2 cos α
p

⎛      ⎞
    
 F α, L sin α G α, L cos α
⎝α   2   2  ⎠ H̄ m sin α+cos
1 mp
1 (α) = − cos α H mp sin α+cos α − sin α
c
 α
 F α, − 2 sin α G α, − 2 cos α
L L p

⎛     ⎞
    
 F α, L sin α G α, − 2L cos α
= ⎝α   2    ⎠ H̄ m sin α+cos
1 mp
− cos α H̄ mp sin α+cos
(α) α − sin α
d
1  α
 F α, − 2 sin α G α, 2 cos α
L L p

⎛     ⎞
    
 F α, − L sin α G α, − 2L cos α
⎝α   2    ⎠ H m sin α+cos
1 mp
2 (α) = − cos α H mp sin α+cos α − sin α
a
 α
 F α, 2 sin α
L L
G α, 2 cos α
p

⎛      ⎞
    
 F α, − L sin α G α, 2L cos α
= ⎝α   2    ⎠ H m sin α+cos
1 mp
− cos α H̄ mp sin α+cos
(α) α − sin α
b
2  α
 F α, 2 sin α G α, − 2 cos α
L L p

⎛     ⎞
    
 F α, L sin α G α, − 2L cos α
⎝α   2    ⎠ H̄ m sin α+cos
1 mp
2 (α) = − cos α H mp sin α+cos α − sin α
c
 α
 F α, − 2L sin α G α, 2L cos α p

⎛      ⎞
    
 F α, L sin α G α, L cos α
= ⎝α   2   2  ⎠ H̄ m sin α+cos
1 mp
− cos α H̄ mp sin α+cos
(α) α − sin α
d
2  α
 F α, − 2 sin α G α, − 2 cos α
L L p

⎛      ⎞
    
 F α, L sin α G α, L cos α
⎝α   2   2  ⎠ H m sin α+cos
1 mp
3 (α) = − cos α H mp sin α+cos α − sin α
a
 α
 F α, − 2L sin α G α, − 2L cos α p

⎛     ⎞
    
 F α, L sin α G α, − 2L cos α
= ⎝α   2    ⎠ H m sin α+cos
1 mp
− cos α H̄ mp sin α+cos
(α) α − sin α
b
3  α
 F α, − 2 sin α G α, 2 cos α
L L p

⎛      ⎞
    
 F α, − L sin α G α, 2L cos α
⎝α   2    ⎠ H̄ m sin α+cos
1 mp
3 (α) = − cos α H mp sin α+cos α − sin α
c
 α
 F α, 2L sin α G α, − 2L cos α p

⎛     ⎞
    
 F α, − L sin α G α, − 2L cos α
= ⎝α   2    ⎠ H̄ m sin α+cos
1 mp
− cos α H̄ mp sin α+cos
(α) α − sin α
d
3  α
 F α, 2 sin α
L L
G α, 2 cos α
p

⎛     ⎞
    
 F α, L sin α G α, − 2L cos α
⎝α   2    ⎠ H m sin α+cos
1 mp
4 (α) = − cos α H mp sin α+cos α − sin α
a
 α
 F α, − 2L sin α G α, 2L cos α p

⎛      ⎞
    
 F α, L sin α G α, 2L cos α
= ⎝α   2    ⎠ H m sin α+cos
1 mp
− cos α H̄ mp sin α+cos
(α) α − sin α
b
4  α
 F α, − 2 sin α G α, − 2 cos α
L L p

⎛     ⎞
    
 F α, − L sin α G α, − 2L cos α
⎝α   2    ⎠ H̄ m sin α+cos
1 mp
4 (α) = − cos α H mp sin α+cos α − sin α
c
 α
 F α, 2L sin α G α, 2L cos α
p

⎛      ⎞
    
 F α, − L sin α G α, 2L cos α
= ⎝α   2    ⎠ H̄ m sin α+cos
1 mp
− cos α H̄ mp sin α+cos
(α) α − sin α
d
4  α
 F α, 2 sin α G α, − 2 cos α
L L p

Remark 2 By comparing Theorems 1 and 2 against metasurfaces in wireless networks. This outcome is deter-
Theorem 3, we observe that the probability of being a mined by the assumption that the metasurfaces can modify
reflector highly depends on the length of the typical object the angle of reflection regardless of their length. The anal-
if is it not coated with a reconfigurable metasurfaces, while ysis of the impact of the constraints imposed by the size of
it is independent of it if it is coated with a reconfigurable the metasurface on its capability of obtaining a given set of
metasurface. This is a major benefit of using reconfigurable angles of reflection as a function of the angle of incidence is
Di Renzo and Song EURASIP Journal on Wireless Communications and Networking (2019) 2019:99 Page 9 of 15

an open but very important research issue, which is left to In Figs. 8 and 9, we depict the probability that the typ-
future research.  ical object is as a reflector as a function of the location
of the transmitter and for a fixed length of the typical
6 Numerical results and discussion: validation object. Once again, the proposed analytical frameworks
against Monte Carlo simulations are accurate, and the upper-bound in Theorem 3 is suffi-
The aim of this section is to validate the analytical frame- ciently accurate for the considered setups. Especially for
works developed in the previous sections against Monte small-size objects, we observe the large gains that employ-
Carlo simulations and to study the potential of using ing metasurfaces bring about. Even small-size objects can
reconfigurable metasurfaces in wireless networks. The provide one with a relatively high probability of being a
results are illustrated either as a function of the length, L, reflector, which is useful information in order to reduce
of the typical object or as a function of the locations of the the deployment cost of the metasurfaces over large-size
transmitter and receiver. The simulation setup is detailed environmental objects. Small-size metasurfaces, in fact,
in the caption of each figure. may be moved along large-size surfaces, and their loca-
In Fig. 7, we validate the proposed mathematical frame- tion may be optimized in order to optimize the system
works, against Monte Carlo simulations, as a function of performance.
the length, L, of the typical object. The results depicted in
Fig. 7 confirm the good accuracy of the proposed analyti- 7 Conclusion and discussion
cal approach. More importantly, we observe the large gain In this paper, we have proposed the first analytical
that the presence of metasurfaces bring about: Especially approach that provides one with the probability that a ran-
for objects of small length, the presence of reconfigurable dom object coated with reconfigurable metasurfaces acts
metasurfaces increases the probability of the typical object as a reflector, and have compared it against the conven-
to be a reflector significantly. This is expected to bring tional setup in which the object is not coated with recon-
major gains in terms of signal strength of the received sig- figurable metasurfaces. This result has been obtained
nal thanks to the reflection generated by the randomly by modeling the environmental objects with a modified
distributed reflectors. The presence of multiple reflec- random line process with fixed length, and random orien-
tors, however, may also increase the level of interference. tations and locations. Our proposed analytical approach
Therefore, the optimization of wireless networks in the allows us to prove that the probability that an object is a
presence of reconfigurable metasurfaces is a challenging reflector does not depend on the length of the object if it
and open research issue. Figure 7, in addition, confirms is coated with metasurfaces, while it strongly depends on
the main finding in Remark 2. it if the Snell’s law of reflection needs to be applied. The

Fig. 7 Probability of being a reflector versus the length of the object. Setup: Rnet = 30 m, location of the transmitter (0, 3), location of the receiver
(20, 20)
Di Renzo and Song EURASIP Journal on Wireless Communications and Networking (2019) 2019:99 Page 10 of 15

Fig. 8 Probability of being a reflector versus the horizontal location, xTx , of the transmitter. Setup: Rnet = 30 m, vertical location of the transmitter
yTx = 3, location of the receiver (0, 0), length of the object L = 5 m

reason of this major difference in system performance lies wireless networks, and to develop a general analytical
in the fact that the angles of incidence and reflection need approach for understanding the ultimate performance
to be the same according to the Snell’s law of reflection. limits, and to identify design guidelines for system opti-
In spite of the novelty and contribution of the present mization. For example, the performance trends are based
paper, it constitutes only a first attempt to quantify the on the assumption that, for any angle of incidence, an
potential of reconfigurable metasurfaces in large-scale arbitrary angle of reflection can be synthetized. Due to

Fig. 9 Probability of being a reflector versus the horizontal location, xTx , of the transmitter. Setup: Rnet = 30 m, vertical location of the transmitter
yTx = 3, location of the receiver (0, 0), length of the object L = 20 m
Di Renzo and Song EURASIP Journal on Wireless Communications and Networking (2019) 2019:99 Page 11 of 15

practical constraints on implementing metasurfaces, only

a finite subset of angles may be allowed, which needs
to account for the concept of field-of-view of the meta-
surfaces. Also, the analytical models and the simulation
results have been obtained by using ray tracing assump-
tions, and ignore, e.g., the radiation pattern of the meta-
surfaces, and near-field effects. A major step is needed to
obtain tractable analytical expressions of relevant perfor-
mance metrics that are suitable to unveil scaling laws, are
amenable for optimization, and account for different func-
tions applied by the metasurfaces (not just reflections).

Appendix 1: Proof of Lemma 3

Let us rewrite (6) and (9) in the following standard forms:
mx − y + z = 0
(18)
mp − y + zp = 0
where their slopes m and mp are assumed to be non-zero.
With this formulation, the directional vector of each line
Fig. 10 Position of a point with respect to a line
is as follows:
   
m mp
= , p= (19)
−1 −1
The two lines in (18) are perpendicular to each other if
their directional vectors are orthogonal: Therefore, the following conclusions can be drawn:
T
p =0 (20) • If the point A is above the line, then it must be
y1 − y > 0. From Eq. (23), we evince that y1 − y > 0
which implies that the following identity need to be ful-
filled mmp = −1. if ax1 +by
b
1 +c
> 0. This, in turn, corresponds to the
following: i) either ax1 + by1 + c > 0 and b > 0 or ii)
The coordinates of the mid-point ofthe line segment ax1 + by1 + c < 0 and b < 0.
 • If the point A is below the line, then it must be
Rx yTx +yRx
between the transmitter and receiver is xTx +x , ,
2 2 y1 − y < 0. From Eq. (23), we evince that y1 − y < 0
which is located on the line y = mp x + zp . By substituting if ax1 +by 1 +c
< 0. This, in turn, corresponds to the
b
this latter point in (18), zp turns out to be the following: following: i) either ax1 + by1 + c < 0 and b > 0 or ii)
1 1 ax1 + by1 + c > 0 and b < 0.
zp = (xTx + xRx ) + (yTx + yRx ) (21)
2m 2
This concludes the proof.
This concludes the proof.
Appendix 3: Proof of Theorem 1
Appendix 2: Proof of Lemma 5 From (10) in Lemma 4, the probability of the complement
Let A (x1 , y1 ) be a point that does not lie on a generic line of Event 1 can be formulated as follows:
ax + by + c = 0. Let us draw the perpendicular line from
A (x1 , y1 ) to the horizontal axis. This perpendicular line
intersect the generic line ax + by + c = 0 in the point # $
  min (xTx , xRx )  x∗  max (xTx , xRx )
P (x1 , y), as shown in Fig. 10. By construction, x1 is the Pr Event 1 = Pr
abscissa of the point P. Let y denote the ordinate of P. ∩ min (yTx , yRx )  y∗  max (yTx , yRx )
Since P is on the line ax + by + c = 0, then the following # $
p−z sin α
min (xTx , xRx )  m sin α+cos α  max (xTx , xRx )
equation needs to be satisfied: = Pr p−z sin α
∩ min (yTx , yRx )  m m sin α+cos α + z  max (yTx , yRx )
ax1 + c
ax1 + by + c = 0 ⇒ y = − (22) (24)
b
Let us compute the difference y1 − y, as follows:
 
ax1 + c ax1 + by1 + c Based on the sign of m sin α + cos α and m, four cases
y1 −y = y1 − − ⇒ y1 −y = (23)
b b can be identified.
Di Renzo and Song EURASIP Journal on Wireless Communications and Networking (2019) 2019:99 Page 12 of 15

Case 1 Case 4
If m sin α + cos α  0 and m  0, we obtain the following: If m sin α + cos α < 0 and m < 0, we have the following:

⎧ ⎫
⎧
   
⎫   ⎨ min x , x   p−z sin α  max x , x  ⎬
  ⎪
⎨ min x , x p−z sin α ⎪
⎬ Tx Rx m sin α+cos α Tx Rx
Tx Rx  m sin α+cos α  max xTx , xRx Pr Event 1 = Pr
Pr Event 1 = Pr     ⎩ ∩ min y , y   m p−z sin α + z  max y , y  ⎭
⎩ ∩ min y , yRx  m p−z sin α + z  max y , yRx ⎪
⎪ ⎭ Tx Rx m sin α+cos α Tx Rx
Tx m sin α+cos α Tx #  δ $
⎧ ⎫ 1 π   2  
⎪       ⎪ = θ4 α, xTx , xRx , yTx , yRx dα + θ4 α, xTx , xRx , yTx , yRx dα
(a) ⎨ f α, min xTx , xRx  υ  f α, max xTx , xRx ⎬
2π δ1 π
= Pr       ⎪
⎪
⎩ ∩g α, min yTx , yRx  υ  g α, max yTx , yRx ⎭  δ
1 2  
 2π minf α,maxx ,x ,g α,maxy ,y ,1 = θ4 α, xTx , xRx , yTx , yRx dα
1 Tx Rx Tx Rx 2π δ1
=         fυ (υ)×H(m sin α+cos α)H(m)
2π 0 max f α,min x ,xRx ,g α,min y ,yRx ,0
⎛
Tx Tx
⎞
(28)
       
⎜ min f α, max xTx , xRx , g α, max yTx , yRx , 1 ⎟
×H⎝         ⎠ dυdα
− max f α, min xTx , xRx , g α, min yTx , yRx , 0
#  2π $ This concludes the proof.
(b) 1 δ1    
= θ1 α, xTx , xRx , yTx , yRx dα + θ1 α, xTx , xRx , yTx , yRx dα
2π 0 δ2

(25) Appendix 4: Proof of Theorem 2

The proof is based on Lemma 5. In particular, the follow-
ing cases need to be examined.
√
where (a) follows from p = Rnet u = Rnet υ and (b) fol- • Case 1: The location of the transmitter (xTx , yTx ) is
lows by computing the integral with respect to υ, whose above the line x cos α + y sin α − p = 0 given
probability density function is fυ (υ) = 2υ, since u is sin α > 0.
uniformly distributed in [0, 1]. • Case 2: The location of the transmitter (xTx , yTx ) is
The integration limits in (b) are determined from the above the line x cos α + y sin α − p = 0 given
conditions m sin α + cos α  0, which implies
 0  α  δ1 sin α < 0.
√
and δ2  α  2π, where δ1 = 2tan−1 m + 1 + m2 • Case 3: The location of the receiver (xRx , yRx ) is above
 √  the line x cos α + y sin α − p = 0 given sin α > 0.
and δ2 = 2π + 2tan−1 m − 1 + m2 . • Case 4: The location of the receiver (xRx , yRx ) is above
The other three case studies can be obtained by using the line x cos α + y sin α − p = 0 given sin α < 0.
the same approach as for Case 1. Thus, the details are • Case 5: The location of the transmitter (xTx , yTx ) is
omitted and only the final result is reported. below the line x cos α + y sin α − p = 0 given
sin α > 0.
Case 2 • Case 6: The location of the transmitter (xTx , yTx ) is
If m sin α + cos α  0 and m < 0, we have the following: below the line x cos α + y sin α − p = 0 given
sin α < 0.
• Case 7: The location of the receiver (xRx , yRx ) is below
⎧ ⎫
  ⎨ min x , x   p−z sin α  max x , x  ⎬ the line x cos α + y sin α − p = 0 given sin α > 0.
Tx Rx m sin α+cos α Tx Rx
Pr Event 1 = Pr 
⎩ ∩ min y , y  p−z sin α   ⎭ • Case 8: The location of the receiver (xRx , yRx ) is below
Tx Rx  m m sin α+cos α + z  max yTx , yRx
#
δ1  $ the line x cos α + y sin α − p = 0 given sin α < 0.
1   2π  
= θ2 α, xTx , xRx , yTx , yRx dα + θ2 α, xTx , xRx , yTx , yRx dα
2π 0 δ2
From (3), the probability of Event 1 can be formulated as
(26) follows:

Case 3 # $
If m sin α + cos α < 0 and m  0, we have the following: {[Tx is above the line] ∩ [Rx is above the line]}
Pr {Event 1} = Pr
∪ {[Tx is below the line] ∩ [Rx is below the line]}
# $
[Case 1 ∪ Case 2] ∩ [Case 3 ∪ Case 4]
⎧ ⎫ = Pr
  ⎨ min x , x   p−z sin α  max x , x  ⎬ ∪ [Case 5 ∪ Case 6] ∩ [Case 7 ∪ Case 8]
Pr Event 1 = Pr
Tx Rx m sin α+cos α Tx Rx
⎩ ∩ min y , y   m p−z sin α + z  max y , y  ⎭
Tx Rx m sin α+cos α Tx Rx = Pr {[Case 1 ∪ Case 2] ∩ [Case 3 ∪ Case 4]}
#  δ $
π    
=
1
θ3 α, xTx , xRx , yTx , yRx dα +
2
θ3 α, xTx , xRx , yTx , yRx dα + Pr {[Case 5 ∪ Case 6] ∩ [Case 7 ∪ Case 8]}
2π δ1 π
 δ = Pr {Case 1 ∩ Case 3} + Pr {Case 2 ∩ Case 4}
1 2  
= θ3 α, xTx , xRx , yTx , yRx dα
2π δ1 + Pr {Case 5 ∩ Case 7} + Pr {Case 6 ∩ Case 8}
(27) (29)
Di Renzo and Song EURASIP Journal on Wireless Communications and Networking (2019) 2019:99 Page 13 of 15

Therefore, four probabilities need to be computed. Let 7.1 Case 1

us start with the first one: If xend1 > xend2 and yend1 > yend2 , which implies sin α < 0
and cos α > 0, we obtain the following:

Pr {Case 1 ∩ Case3} Pr {Event 2}


    p − zp sin α
= Pr xTx cos α+yTx sin α−p > 0∩ sin α > 0 ∩ xRx cos α + yRx sin α − p > 0 ∩ sin α > 0 = Pr xend2 
mp sin α + cos α
&   '  
= Pr xTx cos α + yTx sin α − p > 0 ∩ xRx cos α + yRx sin α − p > 0 ∩ sin α > 0 p − zp sin α
      xend1 ∩ yend2  mp +zp  yend1
(a) x cos α + yTx sin α x cos α + yRx sin α mp sin α + cos α
= Pr υ < Tx ∩ υ < Rx ∩ sin α > 0 ⎧ ⎫
Rnet Rnet  
  ⎪
⎨
L
sin α +
zp sin α

1 L
− cos α p  − sin α +
zp sin α
∩⎪
⎬
 π  min xTx cos α+yTx sin α , xRx cos α+yRx sin α ,1
2 mp sin α + cos α mp sin α + cos α 2 mp sin α + cos α
= Pr
1 Rnet Rnet ⎪
⎩ − L cos α +
  ⎪
= fυ (υ) − zp ⎭
mp zp sin α mp L mp zp sin α
2π 0 0 − zp  − sin α p  cos α +
2 mp sin α + cos α mp sin α + cos α 2 mp sin α + cos α
  
xTx cos α + yTx sin α xRx cos α + yRx sin α
× H min
Rnet
,
Rnet
, 1 dυdα (33)
(b) 1
 π
 
 
1
α − cos α and
= ρ α, xTx , yTx , xRx , yRx dα
2π 0 1 Depending on the sign of mp sin α+cos
 
(30) mp
mp sin α+cos α − sin α , four sub-cases need to be studied.

√ 
Case 1-a 
where (a) follows from p = Rnet u = Rnet υ and (b) 1
If mp sin α+cos α − cos α > 0 and
follows by solving the integral with respect to υ whose  
mp
probability density function is fυ (υ) = 2υ, since u is mp sin α+cos α − sin α > 0, we have the following:
uniformly distributed in [0, 1].
# $
p − zp sin α p − zp sin α
Pr {Event 2}=Pr xend2   xend1 ∩ yend2 mp +zp  yend1
mp sin α + cos α mp sin α + cos α
By using a similar approach, we we obtain the following         
(a) L L L L
results: = Pr F α, sin α  υ  F α, − sin α ∩ G α, − cos α  υ  G α, cos α
2 2 2 2
      ( )
 2π  min F α,− L sin α ,F α,− L sin α ,1
1 2   2 1
=     fυ (υ)H − cos α
 2π 2π 0 L L
max F α, 2 sin α ,G α,− 2 cos α ,0 mp sin α + cos α
1 ( )
Pr {Case 2 ∩ Case 4} = ρ2 (α, xTx , yTx , xRx , yRx )dα H
mp
− sin α
2π π mp sin α + cos α
 π ⎛      ⎞
1 L L
Pr {Case 5 ∩ Case 7} = ρ2 (α, xTx , yTx , xRx , yRx )dα ⎜ min F α, − 2 sin α , F α, − 2 sin α , 1 ⎟
⎜ ⎟
2π 0 ×H⎜
⎝
     ⎟ H̄ (sin α) H (cos α) dυdα
⎠
 2π L
− max F α, sin α , G α, − cos α , 0
L
1 2 2
Pr {Case 6 ∩ Case 8} = ρ1 (α, xTx , yTx , xRx , yRx )dα (b) 1
 2π
1
 2π
2π π =
2π 0
a a
1 (α)H̄ (sin α) H (cos α) dα = 2π 3π 1 (α)dα
2
(31)
(34)
√
where (a) follows from p = Rnet u = Rnet υ and (b)
This concludes the proof. follows by solving the integral with respect to υ whose
probability density function is fυ (υ) = 2υ, since u is
Appendix 5: Proof of Proposition 1 uniformly distributed in [0, 1].
From (8) in Lemma 2 and (11) in Lemma 4, the probability
of Event 2 can be formulated as follows: By using a similar approach, we can study the remaining
three sub-cases.
# $
min (xend1 , xend2 )  x∗  max (xend1 , xend2 )
Pr {Event 2} = Pr
∩ min (yend1 , yend2 )  y∗  max (yend1 , yend2 )  1-b
Case   
⎧ ⎫ 1 mp
⎨ min (xend1 , xend2 )  p−zp sin α  max (xend1 , xend2 ) ⎬ If mp sin α+cos α −cosα > 0 and mp sin α+cos α −sinα < 0,
mp sin α+cos α
= Pr p−zp sin α we have the following:
⎩ ∩ min (yend1 , yend2 )  mp + zp  max (yend1 , yend2 ) ⎭
mp sin α+cos α

(32)  2π
1
Pr {Event 2} = b
1 (α)H̄ (sin α) H (cos α) dα
2π 0
 2π
1 b
In order to compute this probability, we need to examine = 1 (α)dα
four cases depending on the relationship between xend1 2π 3π2
and xend2 , as well as yend1 and yend2 . (35)
Di Renzo and Song EURASIP Journal on Wireless Communications and Networking (2019) 2019:99 Page 14 of 15

 1-c
Case    Case 4
1 mp
If mp sin α+cos α −cos α < 0 and mp sin α+cos α −sinα > 0, If xend1 < xend2 , and yend1 < yend2 , which implies sin α >
we have the following: 0 and cos α < 0, we have the following:
 2π  π
1 1
Pr {Event2} = c
1 (α)H̄ (sin α) H (cos α) dα Pr {Event 2} =
2
4 (α)dα
2π 0 2π 0
 2π (36)  π
1 c 1
= 1 (α)dα = a
4 (α) +
b
4 (α) +
c
4 (α) +
d
4 (α) dα
2π 3π2 2π π2
(41)
 1-d
Case   
1 mp
If mp sin α+cos α −cos α < 0 and mp sin α+cos α −sin α < 0, This concludes the proof.
we have the following:
Funding
 2π Not applicable.
1
Pr {Event 2} = d
1 (α)H̄ (sin α) H (cos α) dα
2π 0 Availability of data and materials
 2π Data sharing is not applicable to this article as no datasets were generated or
1 d analysed during the current study. The paper is built upon mathematical
= 1 (α)dα analysis.
2π 3π2
(37) Authors’ contributions
The authors declare that they have equally contributed to the paper. Both
Therefore, eventually, Pr {Event 2} can be formulated as authors read and approved the final manuscript.

follows: Competing interests

The authors declare that they have no competing interests.
 2π
1
Pr {Event 2} = 1 (α)dα Publisher’s Note
2π 3π 2 Springer Nature remains neutral with regard to jurisdictional claims in

1 2π  a b c d
published maps and institutional affiliations.
= 1 (α) + 1 (α) + 1 (α) + 1 (α) dα
2π 3π2 Publishers Note
Springer Nature remains neutral with regard to jurisdictional claims in
(38)
published maps and institutional affiliations.

By using a similar line of thought, the remaining three Received: 4 January 2019 Accepted: 19 March 2019
cases can be studied. The final result is reported in the
following sections.
References
1. P. Hu, P. Zhang, M. Rostami, D. Ganesan. Braidio: an integrated
active-passive radio for mobile devices with asymmetric energy budgets
Case 2 (ACM SIGCOMM (Florianopolis, Brazil, 2016)
If xend1 > xend2 , and yend1 < yend2 , which implies sin α < 2. C. Liaskos, S. Nie, A. Tsioliaridou, et al., in 2018 IEEE 19th International
Symposium on “A World of Wireless, Mobile and Multimedia Networks”
0 and cos α < 0, we obtain the following: (WoWMoM). Realizing wireless communication through software-defined
hypersurface environments (IEEE, 2018), pp. 14–15
 3π
3. 5GPPP, Vision on Software Networks and 5G SN WG, (2017)
1 2
Pr {Event 2} = 2 (α)dα 4. C. Liaskos, S. Nie, A. Tsioliaridou, A. Pitsillides, S. Ioannidis, I. F. Akyildiz, A
2π π new wireless communication paradigm through software-controlled
 3π  metasurfaces. IEEE Commun. Mag. 56(9), 162–169 (2018)
1 2
a b c d
= 2 (α) + 2 (α) + 2 (α) + 2 (α) dα 5. L. Subrt, P. Pechac, in 2012 6th European Conference on Antennas and
2π π Propagation (EUCAP). Controlling propagation environments using
(39) intelligent walls (IEEE, 2012), pp. 1–5
6. L. Subrt, P. Pechac, Intelligent walls as autonomous parts of smart indoor
environments. IET Commun. 6(8), 1004–1010 (2012)
Case 3 7. X. Tan, Z. Sun, J. M. Jornet, et al., in 2016 IEEE International Conference on
If xend1 < xend2 , and yend1 > yend2 , which implies sin α > Communications (ICC). Increasing indoor spectrum sharing capacity using
smart reflect-array (IEEE, 2016), pp. 1–6
0 and cos α > 0, we have the following: 8. O. Abari, D. Bharadia, A. Duffield, et al., in 14th USENIX Symposium on
Networked Systems Design and Implementation (NSDI 17). Enabling
 π
high-quality untethered virtual reality, (2017), pp. 531–544
1 2
Pr {Event 2} = 3 (α)dα 9. A. Welkie, L. Shangguan, J. Gummeson, et al., in Proceedings of the 16th
2π 0 ACM Workshop on Hot Topics in Networks. Programmable radio
 π 
1 2 environments for smart spaces (ACM, 2017), pp. 36–42
a b c d
= 3 (α) + 3 (α) + 3 (α) + 3 (α) dα 10. R. Chandra, K. Winstein, in ACM workshop on hot topics in networks.
2π 0 Programmable Radio Environments for Smart Spaces - HotNets-XVI
(40) Dialogue, (Palo Alto, 2017)
Di Renzo and Song EURASIP Journal on Wireless Communications and Networking (2019) 2019:99 Page 15 of 15

11. S. Hu, F. Rusek, O. Edfors, Beyond Massive MIMO: The potential of data
transmission with large intelligent surfaces. IEEE Trans. Sig. Process.
66(10), 2746–2758 (2018)
12. X. Tan, Z. Sun, D. Koutsonikolas, et al., in IEEE INFOCOM 2018-IEEE Conference
on Computer Communications. Enabling indoor mobile millimeter-wave
networks based on smart reflect-arrays (IEEE, 2018), pp. 270–278
13. C. Liaskos, A. Tsioliaridou, A. Pitsillides, S. Ioannidis, I. F. Akyildiz, Using any
surface to realize a new paradigm for wireless communications.
Commun. ACM. 61(11), 30–33 (2018)
14. C. Liaskos, A. Tsioliaridou, S. Ioannidis, in 2018 IEEE 23rd International
Workshop on Computer Aided Modeling and Design of Communication Links
and Networks (CAMAD). Towards a Circular Economy via Intelligent
Metamaterials (IEEE, 2018), pp. 1–6
15. N. Yu, P. Genevet, M. A. Kats, F. Aieta, J.-P. Tetienne, F. Capasso, Z. Gaburro,
Light Propagation with Phase Discontinuities: Generalized Laws of
Reflection and Refraction. Science. 334(6054), 333–337 (2011)
16. C. L. Holloway, E. F. Kuester, J. A. Gordon, J. O’Hara, J. Booth, D. R. Smith, An
overview of the theory and applications of metasurfaces: the
two-dimensional equivalents of metamaterials. IEEE Antennas Propag.
Mag. 54(2), 10–35 (2012)
17. L. La Spada, Metamaterials for Advanced Sensing Platforms. Res. J. Opt.
Photonics. 7, 14–16 (2017)
18. T. Nakanishi, T. Otani, Y. Tamayama, et al., Storage of electromagnetic
waves in a metamaterial that mimics electromagnetically induced
transparency. Phys. Rev. B. 87(16), 161110-1–161110-4 (2013)
19. A. Silva, F. Monticone, G. Castaldi, et al., Performing mathematical
operations with metamaterials. Science. 343(6167), 160–163 (2014)
20. H2020 VISORSURF project, A hardware platform for software-driven
functional metasurfaces. http://www.visorsurf.eu/.
21. T. Bai, R. Vaze, R. W. Heath Jr, Analysis of blockage effects on urban cellular
networks. IEEE Trans. Wirel. Commun. 13(9), 5070–5083 (2014)
22. J. Lee, F. Baccelli, in IEEE International Conference on Computer
Communications. On the Effect of Shadowing Correlation on Wireless
Network Performance, (Honolulu, 2018)
23. J. G. Andrews, F. Baccelli, R. K. Ganti, A tractable approach to coverage and
rate in cellular networks. IEEE Trans. Commun. 59(11), 3122–3134 (2011)
24. M. Di Renzo, A. Guidotti, G. E. Corazza, Average rate of downlink
heterogeneous cellular networks over generalized fading channels – a
stochastic geometry approach. IEEE Trans. Commun. 61(7), 3050–3071
(2013)
25. M. Di Renzo, Stochastic geometry modeling analysis of multi-tier
millimeter wave cellular networks. IEEE Trans. Wirel. Commun. 14(9),
5038—5057 (2015)
26. W. Lu, M. Di Renzo, in Proceedings of the 18th ACM International Conference
on Modeling, Analysis and Simulation of Wireless and Mobile Systems.
Stochastic geometry modeling of cellular networks: Analysis, simulation
and experimental validation (ACM, 2015), pp. 179–188
27. M. Di Renzo, W. Lu, P. Guan, The intensity matching approach: a tractable
stochastic geometry approximation to system-level analysis of cellular
networks. IEEE Trans. Wirel. Commun. 15(9), 5963—5983 (2016)
28. M. Di Renzo, A. Zappone, T. T. Lam, M. Debbah, System-level modeling
and optimization of the energy efficiency in cellular networks - a
stochastic geometry framework. IEEE Trans. Wirel. Commun. 17(4),
2539—2556 (2018)
29. M. Di Renzo, S. Wang, X. Xi, Modeling and analysis of cellular networks by
using inhomogeneous poisson point processes. IEEE Trans. Wirel.
Commun. 17(8), 5162–5182 (2018)
30. M. Di Renzo, T. T. Lam, A. Zappone, M. Debbah, A tractable closed-form
expression of the coverage probability in poisson cellular networks. IEEE
Wirel. Commun. Lett. 8(1), 249–252 (2018). https://doi.org/10.1109/LWC.
2018.2868753
31. M. Di Renzo, A. Zappone, T. T. Lam, M. Debbah. Spectral-energy efficiency
Pareto front in cellular networks: a stochastic geometry framework, vol. 8,
(2018), pp. 424–427. https://doi.org/10.1109/LWC.2018.2874642
32. A. Narayanan, S. V. Sreejith, R. K. Ganti, in GLOBECOM 2017-2017 IEEE Global
Communications Conference. Coverage Analysis in Millimeter Wave
Cellular Networks with Reflections (IEEE, 2017), pp. 1–6
33. M. Di Renzo, W. Lu, System-level analysis and optimization of cellular
networks with simultaneous wireless information and power transfer:
Stochastic geometry modeling. IEEE Trans. Veh. Technol. 66(3),
2251–2275 (2017)