See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/228967992

Hidden vs. exposed terminal problem in ad hoc networks

Article · January 2004

CITATIONS READS

92 8,859

4 authors, including:

Aruna Jayasuriya Sylvie Perreau

Central Queensland University Defence Science and Technology Group (DST)
46 PUBLICATIONS   429 CITATIONS    79 PUBLICATIONS   1,662 CITATIONS   

SEE PROFILE SEE PROFILE

Arek Dadej
University of South Australia
48 PUBLICATIONS   524 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

MIMO channel equalization View project

All content following this page was uploaded by Aruna Jayasuriya on 27 October 2015.

The user has requested enhancement of the downloaded file.

 Hidden vs. Exposed Terminal Problem in Ad hoc
Networks
Aruna Jayasuriya, Sylvie Perreau, Arek Dadej, Steven Gordon
Institute for Telecommunications Research
University of South Australia
Mawson Lakes SA 5095, Australia
email: [email protected]

Abstract 802.11 Distributed Coordination Function (DCF)

[3] have gained widespread popularity in ad hoc net-
An ad hoc network is a collection of wireless works. However, all these CSMA/CA based MAC
mobile nodes that are capable of forming a co- protocols suffer from the well known “hidden ter-
operative agreement (network) between them- minal” problem.
selves without requiring any centralized control In wireless networks, it is a commonly ac-
function. Due to their non-reliance on fixed in- cepted practice to use a pre-data control informa-
frastructure, ad hoc networks are gaining popu- tion exchange (virtual medium sensing) to avoid
larity in several networking applications includ- the hidden terminal problem. One such vir-
ing, military, rescue operations and meetings tual sensing mechanism is the 802.11 Request To
and conventions. Generally, protocols used at Send/Clear To Send (RTS/CTS) exchange result-
medium access and physical layers of ad hoc net- ing in nodes getting exclusive access to the channel
works are similar to those used in infrastructure for a well-defined time period. However, the use
based wireless networks. It is a common prac- of RTS/CTS-like schemes introduces the “exposed
tise in infrastructure based wireless networks to terminal problem”1 , where some nodes that heard
use pre-data exchange of control information to the RTS/CTS exchange refrain from transmission
eliminate the “hidden terminal” problem. Al- even though they would not have interfered with
though the use of pre-data exchange of con- any ongoing transmission.
trol information introduces the “exposed termi- The hidden terminal problem was well studied
nals”, it has been shown that throughput perfor- for access points (infrastructure) based networks
mance of infrastructure networks generally im- [5] and it was shown that the introduction of vir-
proves with use of such mechanisms. However, tual sensing (like RTS/CTS) improves their perfor-
ad hoc networks topologies differ significantly mance [1, 2, 4]. Ad hoc networks have gained pop-
from those of access point based (infrastructure) ularity among researchers within the last decade,
wireless networks, hence the effect of hidden ad especially in military and emergency service pro-
exposed terminals in ad hoc networks is differ- vision contexts. Due to technical and commercial
ent than in infrastructure based networks. In reasons, essentially the same distributed MAC layer
this paper we show, through analytical studies protocols used in infrastructure based wireless net-
and simulations, that exposed terminal prob- works have been considered for ad hoc networks.
lem dominates in ad hoc networks, and there- However, ad hoc networks have significantly dif-
fore, contrary to some previously published re- ferent topologies compared with the infrastructure
sults, the use of RTS/CTS handshake may be based networks, leading to the question whether
detrimental to the network’s throughput perfor- RTS/CTS like schemes have the same effect in both
mance. types of networks. In this paper, we discuss the
hidden and exposed terminal problems in terms of
1 Introduction the number of hidden and exposed terminals poten-
tially affecting a communicating pair of nodes in a
Due to the lack of a centralised control entity in large ad hoc network with uniform node density.
ad hoc networks, sharing of wireless bandwidth The rest of this paper is organised as follows.
among ad hoc nodes (medium access control) must Section 2 presents the definitions of hidden and ex-
be organised in a decentralised manner. There- posed terminals in the context of wireless networks.
fore distributed Medium Access Control (MAC) 1 The definition of “exposed terminals” in this paper may
mechanisms such as Carrier Sense Multiple Access differ from other well-known definitions such as the one given
with Collision Avoidance and its’ variants such as in [4]. We define the exposed terminal problem in Section
MACA [1], MACA for Wireless (MACAW) [2] and 2.2.
Furthermore, in this section we formulate the “hid- nal problem. However, this mechanism introduces
den vs. exposed” terminal problem in terms of a new problem termed the exposed terminal prob-
number of hidden and exposed terminals. Section 3 lem. We assume here an RTS/CTS exchange so
derives the number of hidden and exposed terminals that the issue of hidden terminal is addressed. Let
potentially affecting a communicating pair of nodes us consider Figure 2 and assume that node A wants
in a large ad hoc network with uniform node den- to transmit to node B.
sity. This is followed by simulation studies, which
Exposed terminals from C to node pair A/B − Y(x)
show that the network throughput degrades with
RTS/CTS mechanisms in large 802.11 based wire-
less ad hoc networks. In the last section we present
Transmission range of A .D

our concluding remarks as well as suggestions for

further studying the relative effect of hidden and .C
exposed terminals in ad hoc networks.
A . . B
2 Definitions and Problem Formula- x Transmission range of C
tion Area X(x)

In this section, we define the hidden and exposed

Transmission range of B
terminal problems. Furthermore, we present our
framework for comparing the effects of hidden and
exposed terminals in ad hoc network. We assume Figure 2. The exposed terminal problem
that all nodes have the same transmission and re-
ception ranges.
Node A sends an RTS and waits for B to send
2.1 The Hidden Terminal Problem a CTS. Suppose a node D located in area Y (x)
wants to transmit data to node C located in area
Figure 1 illustrates the hidden terminal problem. X(x), and D transmits a RTS to C just before A
Suppose that node A wants to transmit to node sends the RTS to B. After receiving the RTS from
B located at a distance x from A. By only sens- D, C transmits a CTS. This CTS is heard by B
ing the medium, node A will not be able to hear upon which B will enter a backoff period prevent-
transmissions by any node (C) in the dashed area ing B from sending the CTS to A. Therefore, any
denoted by A(x), and will start transmitting, lead- transmission from a node within the area Y to a
ing to collisions at node B. This is the well known node within X(x) will prevent A from transmit-
hidden terminal problem, where the hidden nodes ting data to B, although simultaneous transmis-
are located in the area A(x). sions from area Y (x) to X(x) would not have in-
terfered with transmission from A to B. We define
Transmission range of A the terminals in the region Y (x) as the exposed
terminals for the node pair A/B. In this case, the
Hidden terminals for number of transmissions that could occur between
.C
node pair A/B − A(x)
nodes from area X(x) and nodes from area Y (x)
can be expressed as XY .
A . . B
2.3 Hidden vs. Exposed Terminal
x
Problem
In this paper, we evaluate the relative effects of hid-
den and exposed terminal problems through com-
Transmission range of B parison of the number of such terminals affecting a
given node in an ad hoc network. Let us consider
Figure 1. The hidden terminal problem the following arguments:
• In the case of hidden terminal problem, unsuc-
cessful transmissions result from collisions be-
tween a transmission originated by a node such
2.2 The Exposed Terminal Problem as A which cannot hear the on going transmis-
sions to its corresponding node B. The proba-
RTS/CTS handshake mechanism was introduced to bility of such a collision is proportional to the
wireless MAC layers to eliminate the hidden termi- total number of terminals hidden from A.
 • In the case of exposed terminal, unsuccessful Let us assume a uniform distribution of σ nodes
transmissions result from nodes such as A be- per unit area in the network. Then A(x)σ is the
ing prevented from transmitting, because their number of hidden terminals for a node pair A/B,
corresponding node is unable to send a CTS. when B is located within an annulus dx from dis-
Again such unsuccessful transmissions are pro- tance x from A. Furthermore the probability of
portional to the number of exposed terminals. finding B in this annulus is:
Both these events lead to degradation of a
node’s throughput. 2πxdx
πR2
In sections 2.1 and 2.2, we defined the number Therefore, the average number of terminals, H,
of hidden and exposed terminals for a given node that could be hidden from A, at any time A wants
pair A/B. Therefore, by appropriate integration to transmit is:
we could calculate the average number of hidden
and exposed nodes for a given node. Let us assume Z R
that during time period t, each node has the same 2
H = xA(x)dx
probability, p, of sending a packet to a node in its R2 0
vicinity. Then due to the above arguments, the = 1.3σR2 (2)
ratio of hidden to exposed terminals is proportional
to the relative degradation of a node’s throughput
due to hidden and exposed terminal effects. 3.2 Exposed Terminals
For the hidden terminal case, it is enough for a
In this section, we derive an expression for the num-
nodes in region A(x) (Figure 1) to transmit to any
ber of exposed terminals for a given node in an ad
node in its vicinity. However, in the exposed termi-
hoc network. Let us consider Figure 3:
nal problem node D in region Y (x) should send a
RTS to C for the node B to hear the CTS send by Exposed terminals from C to node pair A/B − Y(x)
C 2 . Therefore the effect of exposed terminals has
to be further weighted by the average probability Transmission range of A .D
of a given node (D) in region Y (x) communicating
with a given node, (C) in region X(x).
.
C
3 Effects of Hidden and Exposed θ
y

Terminals A . . B
α

Transmission range of C
In this section, we derive the number of hidden and Area X(x)
exposed terminals affecting a given node in an ad
hoc network.
Transmission range of B

3.1 Hidden Terminals Figure 3. Exposed terminal calculation

The shaded region (with area A(x) in Figure 1)

shows the number of hidden terminals affecting
node A. By using basic geometric arguments and Using basic geometry, the notations presented
assuming that the transmission range is R, it can in Figure 3, symmetry of the system and the ar-
be shown that: guments presented in Section 3.1 it can be shown
that:
³ x ´ xp
A(x) = πR2 − 2R2 arccos + 4R2 − x2 y yp 2
2R 2 Y (x, y, α) = πR2 −2R2 arccos ( )+ 4R − y 2
(1) 2R 2
Note that A(x) only corresponds to the area (3)
containing nodes hidden from A when A wants to In Figure 4 it can be observed that the expres-
transmit to B. In order to find the total number sion for area Y could take different forms depending
of nodes potentially hidden from A, one needs to on the position of node C within the area X(x).
consider all possible nodes within the transmission If node C is located outside the circle centred
range of node A. around point T , (intersection of transmission ranges
2 There are other nodes in the region A(x) that would
of A and B), the area Y can be expressed by the
send CTS that could be heard by B. These nodes are taken
Equation (3). However if the node C is located out
into account during the integration process over the region side this circle then the area for exposed terminal
X(x). is given by the expression:
 Exposed terminals from C to node pair A/B − Y
In Section 2.3 we explained that for the exposed
Area Y3 terminal case it is necessary for node D in region
Y to communicate with node C in region X (re-
fer to Figure 2) to experience the effect of exposed
. .
T
.C
.
T
terminals. Therefore we should weight the number
of exposed terminals given by (18) by the average
. probability of the event that a given node in the
A
. . B A
. . B
C
region Y will be communicating with a given node
in region X.
Assume each node communicates with all other
nodes in its’ vicinity with the same probability.
Area X(x)
Then the probability of Node D communicating
with C (refer to Figure 2) is:
Figure 4. Various scenarios for exposed terminal
calculations σydαdy
σπR2
Then the average probability of this event is:
Z R Z π−θ Z R
2 y
Pav = 2x dy dα dx
R2 πR2
y yp 0
Z µ
0 r1
µ ¶¶
πR2 −2R2 arccos ( )+ 4R2 − y 2 − Y3 (x, y, α) 2 R
3 2 4 xS
2R 2 =
πR6 0
x S(x − xR ) + 2R arctan
2R2 − x2
dx

= 0.28
Then from Figure 3 the number of exposed ter-
minals can be found by integrating over the region
XY as follows: Where r1 , θ and S are as defined before.

Z R Z π−θ Z R On the other hand if we assume that a given

2 node D has a probability 1 of communicating with
x 2Y (x, y, α) σ 2 y dy dα dx (4)
R2 0 0 r1 given node C, then we can estimate that the num-
ber of average exposed terminals for a given node
Where
is between 0.29σ 2 R4 and 1.03σ 2 R4 .
R - Transmission/reception range of the
As we are interested in the lower bound on the
nodes
number of exposed terminals we can conclude that
the number of exposed terminals affecting a node in
σ - Node density of the network
an ad hoc network is greater than 0.29σ 2 R4 . Figure
x 5 shows the number of hidden terminals and the
θ = arccos ( 2R )
lower bounds on the number of exposed terminals
p for various node densities.
r1 = R2 − x2 sin2 α − x cos α
It can be observed in Figure 5 that for densities
y greater than 4 nodes per transmission neighbour-
Y (x, y, α) = πR2 − 2R2 arccos ( 2R )
y
p hood, the number of exposed terminals dominate
+ 2 4R − y2 2
the number of hidden terminals. Assuming that the
= Y1 (x, y, α) effect (degradation of performance) of hidden and
exposed terminals on the network performance is
or proportional to the number of hidden and exposed
y terminals perceived by a pair of nodes (see Section
Y (x, y, α) = πR2 − 2R2 arccos ( 2R ) 2.3), we can conclude that for most topologies of
y
p
2 2
+ 2 4R − y − Y3 (x, y, α) large ad hoc networks the exposed terminal effect
= Y2 (x, y, α) dominates the hidden terminal effect. It has to be
emphasise here that this result is only applicable
depending on the position of C. to networks large enough, such that the majority
of the nodes have the full influence of exposed ter-
Although the integration presented in equation minals. In other words, the majority of the nodes
(4) is not tractable, a lower bound for this can be should be at least 3 times the transmission range
derived. From this derivation, presented in Ap- away from all the boundaries of the network. In
pendix A, it can be shown that the number of ex- applications of ad hoc networks, such as military
posed terminals for node A, E, is: networks or conventions generally contain a large
number of nodes. Therefore it beneficial to oper-
E > 1.03σ 2 R4 (5) ate these ad hoc network without the RTS/CTS
 30 1600
Without RTS/CTS
With RTS/CTS
1400
25 Exposed
Hidden
1200

Throughput per node (kbps)

Number of terminals

20
1000

15 800

600

10

400

5
200

0
0 4 5 6 7 8 9 10 11 12 13
1 2 3 4 5 6 7 8 9 10
Number of nodes per transmission neighbourhood
Number of nodes per unit area

Figure 5. Variation of hidden and exposed termi- Figure 6. Throughput performance with and with-
nals with node density out RTS/CTS

mechanism to eliminate the performance degrada- simulation results suggest that throughput perfor-
tion caused by the exposed terminals. mance degrades due to use of RTS/CTS like pre-
data handshake mechanisms. This argues strongly
against the use of RTS/CTS in ad hoc networks,
4 Simulations contrary to the widely published results [6, 7, 8].
It is emphasised here that Figure 5 shows the
Network simulations using OPNET network simu- variation of number of hidden and exposed termi-
lator were conducted to further establish the find- nals, not the effect of hidden and exposed terminals
ing presented in section 3. The nodes were dis- on the performance of ad hoc networks. Therefore
tributed in a grid pattern evenly throughout the this study does not take into account all the effects
network. In these simulations the average through- of hidden and exposed terminals on the transmis-
put of nodes in the centre of the network was calcu- sion patters of ad hoc networks. Furthermore we
lated for different node densities with and without do not take into account the effect of data rate or
the RTS/CTS mechanism. A network with station- the condition of the physical channel in to consid-
ary nodes, each communicating only with the neigh- eration. A better method would be to evaluate the
bouring nodes using 802.11 medium access control relative transmission opportunities and the relative
layer and Dynamic Source Routing (DSR) was used success of these transmissions in ad hoc networks
in the simulation. Figure 6 shows the node through- with and without the RTS/CTS scheme, which can
put for different scenarios. take into account parameters such as the data rate
It can be seen in Figure 6 that the through- and conditions of the physical channel. The trans-
put at the centre of the network can be improved mission opportunities in the network in conjunction
on average by 50% without RTS/CTS. This agrees with the probability of packet loss due to collisions
with the analytical findings that the performance can then be directly related to the network through-
is degraded by the exposed terminals compared to put.
the effect of hidden terminals. Therefore, by comb-
ing the results of analytical and simulation studies
it can be concluded that the throughput of wire- Acknowledgments
less ad hoc networks can be improved by not using
handshake mechanisms such as RTS/CTS. Authors wish to thank Mr. Yong Liaw for his sup-
port during the simulation and discussions during
the analytical studies. This work was supported by
5 Conclusions Defence Science Technology Organisation.

In this paper, we evaluated the relative through- References

put performance of ad hoc networks with and with-
out pre-data control schemes, such as RTS/CTS [1] P. Karn, “MACA - a New Channel Access
mechanism in 802.11 system. The analytical and Method for Packet Radio,” in Amateur Radio
 9th Computer Networking Conference, pp. 134–
140, September 1990.
[2] V. Bharghavan, A. Demers, S. Shenker, and
L. Zhang, “MACAW: A Media Access Proto-
col for Wireless LANs,” in ACM SIGCOMM’94, .
T
α2
pp. 212–225, 1994. r2

[3] “ANSI/IEEE Std 802.11 - 1999 Wireless LAN . C

Medium Access Control (MAC) and Physical α1

[4]
(PHY) Specifications,” 1999.
V. Bharghavan, “Performance Evaluation of Al- A
. . B
gorithms for Wireless Medium Access,” in IEEE
International Computer Performance and De-
pendability Symposium, pp. 86–95, September
1998.
[5] L. Kleinrock and F. Tobagi, “Packet Switching
in Radio Channels Part II - the Hidden Node Area X(x)
Problem in Carrier Sense Multiple Access Nodes
and the Busy Tone Solution,” IEEE Transac- Figure 7. Parameters for exposed terminal calcula-
tions on Communications, vol. COM-23, no. 12, tions
pp. 1417–1433, 1975.
[6] Y. Wang and J. J. Garcia-Luna-Aceves, “Col-
lision Avoidance in Multi-Hop Ad Hoc Net-
works,” in IEEE/ACM Intl. Symposium on Now let us consider the following integral E(x):
Modelling, Analysis and Simulation of Com- Z π−θ Z R
E(x) = Y (x, y, α) dα dy
puter and Telecommunication Systems (MAS- 0 r1
COTS’02, 2002. Z α1 Z R Z α2 Z r2
2 2
= Y1 (y, α) σ y dy dα + Y2 (y, α) σ y dy dα
[7] P.Chatzimisios, V. Vitsas, and A. C. Boucou- 0 r1 α1 r1
Z Z Z Z
valas, “Throughput and Delay Analysis of IEEE α2 R
2
π−θ R
2
+ Y1 (y, α) σ y dy dα + Y2 (y, α) σ y dy dα
802.11 Protocol,” in 5th IEEE International α1 r2 α2 r1
Z Z Z Z
Workshop on Networked Appliances, pp. 168– =
π−θ R
2
Y1 (y, α) σ y dy dα −
α2 r2
2
Y3 (y, α) σ y dy dα
174, 2002. 0
Z
r1 α1 r1
π−θ Z R
[8] J. Weinmiller, H. Woesner, J. Ebert, and −
2
Y3 (y, α) σ y dy dα
A. Wolisz, “Analyzing the RTS/CTS Mecha- α2 r1

= B(x) − C(x) − D(x) (6)

nism in the DFWMAC Media Access Protocol
for Wireless LAN’s,” in IFIP TC6 Workshop
Personal Wireless Communications (Wireless Where
Local Access), pp. 117–130, 1995. α1 = π
6

2π x
α2 = 3 − arccos ( 2R )
Appendix A p
r1 = R2 − x2 sin2 α − x cos α
In this appendix we provide the solution for the √
r2 = 4R2 − x2 sin α − x cos α
integration presented in equation (4). This derives
the number of exposed terminals affecting a pair of
communicating nodes in an ad hoc network.
Let us consider A(y), as defined in Equation 7
In figure 7, assume that (α1 , r1 ) and (α2 , R) are
the intersection points for a circle centred around µ ¶ Ã √ !
2 2 y yp 2 π 3
point T and circles centred around points A and B A(y) = πR −2R arccos
2R
+
2
4R − y 2 −
3
+
2
Ry
respectively (refer to Figure 4). Also let us assume (7)
that in the range α1 < α < α2 , y = r2 is the
distance between nodes B and C when C is located The first and second derivatives of A(y) are
distance R away from point T , as shown in Figure given by:
7. Therefore for α1 < α < α2 :
à √ !
p π 3
0
A (y) = 4R2 − y 2 − + R (8)
3 2
If y > r2 Y (x, y, α) = Y1 (x, y, α)
otherwise Y (x, y, α) = Y2 (x, y, α) and
 ³ x ´
00 −y Y3,max (x) = R2 sin (13)
A (y) = p 2R
4R2 − y 2
As now we have found a suitable expression for
Itr
can be shown that in the range y = 0 . . . R, Y3 , we can now evaluate the terms C(x) and D(x)
³ √ ´2
y = 4 − π3 + 23 R = ymax is the only solution in Equation (6). Using Equations (6) and (13) and
the arguments outlined in the previous paragraph,
for A0 (y) = 0 and A00 (ymax ) < 0. Therefore in we can deduce that:
the rangery = 0 . . . R, A(y) has a local maxima at
³ √ ´2
ymax = 4 − π3 + 23 R and A(ymax ) = 0.034. ³ x ´ Z α2 Z r2
C(x) < 2R2 sin σ 2 y dydα
Furthermore A(0) = 0 and A(R) = 0, implying that 2R α1 r1
A(y) is strictly non-negative in the range 0 . . . R. ³ x ´ Z α2
= R2 σ 2 sin r2 − r12 dα (14)
2R α1 2
Therefore,
à Where r1 and r2 are as defined earlier. Using
√ !
π 3 methods used in earlier in this section3 , it can be
Y1 (y, α) > + Ry (9) shown that:
3 2

In this work, our objective is to show that the ¡ x ¢³

R2 σ 2 sin 2R π´
number of exposed terminals for a node is greater C(x) < α2 − [Γ − Θ]
2 6
than the number of hidden terminals for the same = Cupper (x) (15)
node. As the integral (4) is not trivial to solve,
we have decided to estimate a lower bound for (4)
2π x
and show that this lower bound is greater than the Where α2 = 3 − arccos ( 2R )
number of hidden terminals given by Equation (2). ¡√ ¢2
Γ= 4R2 − x2 sin α2 − x cos α2
³p ´2
Using (9) and (6) it can be shown that: Θ= R2 − x2 sin2 α2 − x cos α2

Z Z
B(x) > 2KRσ
2
π−θ R
2
y dy dα
We can now find a lower limit for D(x) by using
0 r1 (13):
Z p
2KRσ 2 π−θ
3 3
= R −( R2 − x2 sin2 α − x cos α) dα
3 0 µ ¶ Z π−θ Z R
2 x 2
(10) D(x) < 2R sin σ y dydα
2R α2 r1
µ ¶ Z π−θ
2 2 x 2
³ = σ R sin R − Λ dα
√ ´ 2R α2
π 3
Where K = 3 + 2 . σ2
µ
x
¶h ³ ´ √ 2 2i
3 2
= sin S 2x − 3R x + Φ + 3R x
4 2R
Using a argument similar to the one used the = Dupper (x) (16)
in previous section, it can be shown that for all
y = y1 . . . R:
Where α2 , θ and r1 are as defined before.
p 3 3
√
3 3
R −( R2 − x2 sin2 α−x cos α) >
(α − (π − θ))(R − (R − x) ) S = ³ 4R2 − x2 ´2
π−θ p
(11) Λ= R2 − x2 sin2 α − x cos α
³ ³ ´ ¡ ¢´
Therefore, by using Equations (10) and (11); Φ = 4R4 arctan 2RxS 2 −x2 − arctan Sx

Having found suitable expression for B(x), C(x)

KRσ 2 and D(x), we can now evaluate the integral (6).
B(x) > (π − θ)(R3 − (R − x)3 ) = Blower (x)
3 Substituting from inequalities (12), (15) and (16);
(12)
E(x) < Blower (x) − Cupper (x) − Dupper (x) (17)
From Figure 4 it can be observed that it is not
trivial to calculate the area Y3 for a given x, y and α. The number of potential exposed terminals for
As we argued earlier, the objective of this study is to the node pair A/B is calculated by weighted inte-
find the lower bound of the integral 4), which allows gration of E(x) from x = 0 . . . R. Therefore, a lower
the substitution of the maximum of Y3 , Y3,max , for 3 By calculating the first and second derivatives of the
Y3 in the integral 6). Furthermore, it can be shown appropriate function it can be shown that the corresponding
that: function is strictly negative in the region α1 < α < α2 .
 View publication stats

limit for the number of exposed terminals could be

estimated by integrating both sides of the Equation
(17):

Z R
2
Elower = Blower (x) − Cupper (x) − Dlower (x) dx
R2 0

= 1.03σ R4 2
(18)