Improving Routing Performance Using Cooperative Spectrum Sensing in Cognitive Radio Networks

Abstract

The traditional fixed spectrum assignment policy, in wireless networks, has led to significant underutilization (both spatially and temporally) of some licensed spectrum bands and crowdedness of unlicensed spectrum bands. These challenges gave birth to the new spectrum utilization paradigm called “Opportunistic Spectrum Access (OSA)”. Networks that operate under this new paradigm are called Cognitive Radio Networks, named after the enabling technology of OSA. The new paradigm allows unlicensed wireless users, also called Secondary Users (SUs), to use licensed spectrum bands as long as they are not in use by their licensed users, also called Primary Users (PUs). One of the most important properties of CRNs is that channel availability changes over time depending on the activity of PUs. Therefore, SUs must be able to detect PU activity on licensed spectrum bands, in order to use those bands for data communication when they are not in use by their PUs. This nature affects many functions in the network including routing. Routing in CRNs is different from traditional network routing, since it requires spectrum availability awareness. Therefore in CRNs, all intermediate SUs must sense channels availability periodically. However, the overall sensing time over a selected route cannot be neglected. In fact, the overall transmission time for SUs along a route is reduced, due to the time spent on the required periodic sensing for these SUs.
In this paper, we introduce a novel Cooperative Spectrum Sensing (CSS) strategy, in which SUs along a selected route cooperate with their neighboring SUs to monitor PUs’ activities. In our proposed strategy, a SU along the route selects a neighboring SU, if exists, to conduct spectrum sensing on its behalf for a particular channel. This selection is based on the required channel sensing time, and the remaining available time of the candidate SU. Simulation results show that the proposed model improves routing performance such that it reduces the overall required sensing time along selected routes, and therefore, the available time that SUs can offer for data transmission is increased. Also, the end-to-end delay and the achieved bottleneck link rate are enhanced.
Copyright © 2016 Praise Worthy Prize - All rights reserved.

Full text

International Review on Computers and Software (I.RE.CO.S.), Vol. 11, N. 10
ISSN 1828-6003 October 2016
Copyright © 2016 Praise Worthy Prize S.r.l. - All rights reserved DOI: 10.15866/irecos.v11i10.10716
923
Improving Routing Performance Using Cooperative Spectrum Sensing
in Cognitive Radio Networks
Sharhabeel H. Alnabelsi1, Ramzi R. Saifan2, Hisham M. Almasaeid3
Abstract – The traditional fixed spectrum assignment policy, in wireless networks, has led to
significant underutilization (both spatially and temporally) of some licensed spectrum bands and
crowdedness of unlicensed spectrum bands. These challenges gave birth to the new spectrum
utilization paradigm called “Opportunistic Spectrum Access (OSA)”. Networks that operate under
this new paradigm are called Cognitive Radio Networks, named after the enabling technology of
OSA. The new paradigm allows unlicensed wireless users, also called Secondary Users (SUs), to
use licensed spectrum bands as long as they are not in use by their licensed users, also called
Primary Users (PUs). One of the most important properties of CRNs is that channel availability
changes over time depending on the activity of PUs. Therefore, SUs must be able to detect PU
activity on licensed spectrum bands, in order to use those bands for data communication when
they are not in use by their PUs.
This nature affects many functions in the network including routing. Routing in CRNs is different
from traditional network routing, since it requires spectrum availability awareness. Therefore in
CRNs, all intermediate SUs must sense channels availability periodically. However, the overall
sensing time over a selected route cannot be neglected. In fact, the overall transmission time for
SUs along a route is reduced, due to the time spent on the required periodic sensing for these SUs.
In this paper, we introduce a novel Cooperative Spectrum Sensing (CSS) strategy, in which SUs
along a selected route cooperate with their neighboring SUs to monitor PUs’ activities. In our
proposed strategy, a SU along the route selects a neighboring SU, if exists, to conduct spectrum
sensing on its behalf for a particular channel. This selection is based on the required channel
sensing time, and the remaining available time of the candidate SU. Simulation results show that
the proposed model improves routing performance such that it reduces the overall required
sensing time along selected routes, and therefore, the available time that SUs can offer for data
transmission is increased. Also, the end-to-end delay and the achieved bottleneck link rate are
enhanced. Copyright © 2016 Praise Worthy Prize S.r.l. - All rights reserved.
Keywords: Cognitive Radio Networks, Dynamic Spectrum Access, Cooperative Spectrum
Sensing, Routing
Nomenclature
CSS Cooperative spectrum sensing
PU Primary User
SU Secondary Users
n Number of SUs along the selected route
g Set of SUs within the selected route
G The set of neighbor SUs for a SU over the
selected route
G
i
A SUi from the set of neighboring SUs for a
SU over the selected route
SUr A SU along the selected route
SUn A neighboring SU for another SU along the
selected route
Τ The maximum remaining time for a SU in G
SUs set
Tcyc The monitoring cycle time


i
idle
T %
SUi idle time percentage
i
sen
T
SUi channel sensing time
i
sw
T
SUi channel switching time
i
busy
T
SUi time utilized for other route transmission
and sensing, if exist
i
x
T
SUi transmission time delay
L Data segment size (Mb)
Ui Utilized percentage of monitoring cycle that
used only for data transmission by SUi
β Bottleneck link rate in a selected route
R Channel bandwidth (Mbps)
RTT Round trip time
I. Introduction
Nowadays, the unlicensed spectrum bands became
crowded, while the licensed spectrum bands generally is
less crowded and temporally and spatially are

S. H. Alnabelsi, R. R. Saifan, H. M. Almasaeid
Copyright © 2016 Praise Worthy Prize S.r.l. - All rights reserved International Review on Computers and Software, Vol. 11, N. 10
924
underutilized. The measurements of Federal
Communications Commission (FCC) on spectrum
utilization have shown that the utilization for some
channel bands could be very low, and the overall
spectrum utilization is less than 15% as reported in [1].
Therefore, a new spectrum access paradigm has been
proposed, namely Cognitive Radio (CR), which
dynamically allows unlicensed users to access the
licensed spectrum, if available [2].
Cognitive radio technology allows the unlicensed
wireless network users; also called Secondary Users
(SUs), to adjust their transceivers frequencies depending
on the availability of licensed frequency bands when they
are not used by their licensees [3].
In this technology, SUs can dynamically and
opportunistically share and access the unused licensed
channels, in order to improve their performance metrics
such as the End-to-End (E2E) delay, throughput, and
service reliability. One important condition in
opportunistic spectrum sharing is SUs must evacuate the
licensed channels when their licensees, also called
Primary Users (PUs), become active again.
Cognitive Radio Networks (CRNs) face many
challenges such as spectrum sensing, management,
mobility, allocation and sharing [4], [5], attacks [6],
interference and collision [7]-[10], routing protection
[11] due to dynamic channels availability, and channels
assignment for SUs [12].
Spectrum sensing is very important for SUs in CRNs,
because all SUs within the PU’s surrounding interference
area and currently using its licensed channel must
evacuate when PU becomes active again, in order to
avoid interference with the PU’s transmission. Note that
it is well known that each SU uses a PU’s channel must
check every some certain time, known as monitoring
cycle in literature, whether the corresponding PU’s
channel becomes active again or not.
In this paper, we propose a Cooperative Spectrum
Sensing (CSS) strategy between SUs along a given
selected route and their neighbor SUs. In our proposed
strategy, for any SU over a route, call it SUr, other SUs
which are within the same geographic area of PU activity
as SUr can be employed to sense the channel availability,
and report sensing results to SUr.
Therefore, in every monitoring cycle, SUr does not
need to sense the licensed channel availability.
Therefore, SUr can use the saved sensing time to
transmit more data instead. Thus, the overall throughput
over the selected route is increased, and the End-to-End
delay is decreased. The rest of the paper is organized as
follows: in Section II, the related work is presented.
Section III explains preliminaries for our work. Section
IV presents our motivation. Section V, we explained our
proposed cooperative spectrum sensing model. Section
VI, the end-to-end delay and throughput enhancement is
discussed. Section VII shows our simulation results.
Finally, conclusions are presented in Section VIII. The
abbreviations meaning that used in this work are shown
in the nomenclature Table.
II. Related Work
Several studies addressed cooperative spectrum
sensing in CRNs targeting different objectives.
Such objectives include increasing sensing accuracy
and achieving energy efficiency or optimizing network
throughput while maintaining a certain level of sensing
accuracy. Next, we review some of these studies.
II.1. Cooperative Spectrum Sensing (CSS)
for Accuracy
A cooperative spectrum sensing mechanism is
proposed in [13] to overcome the multipath fading and
shadowing effect. In their work sensing is constrained by
a limited time and number of SUs which perform
sensing. This work reduces the energy consumption
caused by channels sensing, and computes the number of
SUs that should participate in sensing.
Overcoming the effect of fading and shadowing by
collaborative sensing is also studied in [4]. In [14], the
authors investigated the detection performance of an
energy detector used in cooperative spectrum sensing
taking into consideration both multi-path fading and
shadowing. Using data fusion, the problem is analyzed in
four different scenarios to derive upper bounds on the
detection probability. Using decision fusion, on the other
hand, the exact detection and false alarm probabilities are
derived. In [15], the problem optimizing the ability of
detecting weak primary signals using cooperative
spectrum sensing is investigated. An optimal linear
cooperation framework that uses linear combination of
local statistics from individual SUs is proposed. In [16], a
clustering-based cooperative spectrum sensing method is
proposed. SUs are divided into clusters based on users’
diversity such that the node with the largest channel gain
is selected as the cluster head. Cluster-heads send sensing
results on behalf of cluster members.
The proposed method achieves significant
improvement in the sensing performance by exploiting
the diversity in user selection as reported in the paper.
II.2. Energy Efficient CSS
An energy efficient spectrum sensing method which
depends on the combination of censoring and sleeping
policies is proposed in [17]. It is based on given priori
information about the activity of PUs, and the maximum
allowable false alarm, their results show SUs energy
consumption is reduced. In [18], the authors proposed the
use of a sensor network dedicated for spectrum sensing
in order to aid cognitive radio networks. The sensor
nodes are divided into a set of clusters, and the optimal
activation schedule of these clusters is calculated in order
to minimize energy consumption while meeting
necessary detection and false alarm thresholds.
The problem of energy-efficient spectrum sensing
scheduling with satisfactory PU protection is also studied
in [19].

S. H. Alnabelsi, R. R. Saifan, H. M. Almasaeid
Copyright © 2016 Praise Worthy Prize S.r.l. - All rights reserved International Review on Computers and Software, Vol. 11, N. 10
925
A model that uses the signal-to-noise ratio (SNR) of
the primary signal at various SUs is proposed, which
helps to determine how long each SU must sense each of
the channels, in order to achieve reliable energy efficient
sensing performance. The length of the sensing period
affects the total consumed energy of SUs due to sensing.
Therefore, the authors in [20] proposed a scheme in
which the lengths of sensing periods of various channels
are dynamically adjusted based on profiling PUs
activities. An optimized energy and time schemes for
CSS are proposed in [21]. SUs conduct a short sensing
operation, which is sufficient when the SNR is high or
when the intended PU is inactive. If this first stage is not
sufficient, a second stage of fine spectrum sensing will
be performed to increase the spectrum sensing accuracy.
Thus, the required sensing time and consumed energy
are reduced.
II.3. CSS with Maximized Throughput
A CRN that consists of a single PU and multiple SUs
is used in [22] to study the problem of exploiting
cooperative spectrum sensing to maximize the total
expected system throughput. A Bayesian decision rule
based algorithm was proposed to solve the problem
optimally with a constant time complexity.
A novel capacity-aware cooperative spectrum sensing
method was proposed in [23]. The proposed optimization
method computes the optimal cooperative sensing
parameters like the number of sensed samples, the
number of cooperating nodes, and the used bandwidth on
control channel such that the secondary system capacity
is maximized. In [24], the k-out-of-N fusion rule is used
in cooperative sensing to determine the presence of the
PU. The author studied the problem of finding the pair of
sensing time and the value of k that maximize the SUs’
throughput subject to sufficient protection to PUs.
III. Preliminaries
In this section, we give some preliminary definitions
and highlight some of the assumptions we make in this
work.
Fig. 1. Primary user transmission, sensing, and interference ranges
with radii: r1, r2, r3, respectively
Definition III.1: Sensing time is the amount of time
required for a particular SU to sense the activity of PUs
on a given spectrum channel. This amount may differ
depending on the sensed channel as well as the hardware
specifications of the SUs radio device [25].
Definition III.2: Monitoring Cycle is the amount of
time that passes before the channel must be sensed.
Definition III.3: Secondary user available Time: the
amount of time that is not dedicated for transmission on
the current path (when the SU employed by other paths).
A PU is usually associated with different types of
ranges (depicted in Figure 1) which are:
1. The transmission range: in which another PU can
receive the transmitted data signal, shown with a
radius of r1.
2. The PU sensing range: within this area SUs can sense
the activity of PU with an acceptable accuracy,
shown with a radius of r2.
3. The PU’s interference range: in which SUs cannot
transmit data if the PU is active, shown with a radius
of r3, in order to avoid interference with PU
transmission.
We assume in this work that when a SU wants to
choose another SU for cooperative sensing on a specific
channel, both SUs must be affected by the same group or
set of PU(s) on that channel.
This condition guarantees that the sensing decisions
made by both SUs are consistent and have the same
meaning.
To explain this assumption, Fig. 2 shows an example
scenario, assuming that all PUs operate on the same
channel, only SU2 and SU7 can do cooperative sensing
on each other’s behalf, while other SUs pairs cannot,
e.g.; SU3 cannot do cooperative sensing for SU2, since
SU3 is affected by all PUs transmissions, while SU2 is
affected only by PU2 transmission. Therefore, if SU2
sensed the channel to be busy or idle, then it is also busy
or idle at SU7.
Fig. 2. Selection conditions for a cooperative spectrum sensing SU

S. H. Alnabelsi, R. R. Saifan, H. M. Almasaeid
Copyright © 2016 Praise Worthy Prize S.r.l. - All rights reserved International Review on Computers and Software, Vol. 11, N. 10
926
IV. Motivation
For a given selected route in CRNs, SUs along this
route may have some neighboring SUs which may have
an idle time (where this idle time is enough to sense the
currently used licensed channels, in order to decide
whether their corresponding PUs become active again or
not). In order to have a cooperative spectrum sensing
between a SU along the route and its SU neighbor for a
channel on each other’s behalf (as explained in previous
section using Fig. 2 network scenario), they must be
within the sensing range of the same set of PUs operating
on this channel.
Building on these two observations, we were
motivated to propose the following:
1. For SUs along a given selected route, we propose to
utilize their neighbor SUs’ idle time, if exist, in order
to perform cooperative channel sensing, while SUs
along the route are busy with data transmission. As a
result, the transmission time for SUs along the route
is increased, and therefore, the throughput is
increased.
2. SUs in a network may differ by their required sensing
time which is required for spectrum sensing, since the
needed sensing time is based on SUs channel
conditions, and other factors [25]. Therefore, we are
motivated to select a neighbor SU(s) which needs the
least channel sensing time, and has largest idle time
(such that this idle time is enough for the required
sensing process). This selection factor is formulated
in the objective function, equation (2), as explained in
Section V.
Fig. 3 gives a motivational example. Assume that
PU1, PU2, PU3, PU4, and PU5 are primary users which
have the license to use channels: ch1, ch2, ch3, ch4, and
ch2, respectively. Also, assume that at a given time, the
common available channel between SU1 and SU2 is ch1,
and between SU2 and SU3 is ch2, and between SU3 and
SU4 is ch2, and between SU4 and SU5 is ch3.
The circles in this Figure represent the interference
ranges of PUs. To choose a SU for cooperative sensing,
it must be affected by the same set of PU(s).
For example in Fig. 3, SU3 may select SUC1, because
both are affected by PU2 and PU5 interference ranges,
while SUC2 cannot be selected, since it is affected by
PU2 only.
Fig. 3. An example of given selected route for SUs in CRNs
Notice that when SU3 wants to transmit to SU4 over
ch2, both PU2 and PU5 must be idle, where SUC1 is the
only neighbor SU which is capable to sense whether
these two PUs is active or not, due to the aforementioned
reason. Let’s discuss another motivation, Fig. 3 shows
SU4 has three neighboring SUs (that are affected by the
same PU, called PU3) named S1, S2, and S3. Let’s
assume their available or remaining times are: 10%,
12%, and 15%, respectively, of the monitoring cycle
time, also assume these SUs required sensing times for
the licensed channel are: 50 ms, 50 ms, and 100 ms,
respectively. Also, assume SUs monitoring cycle time =
1000 ms. In our proposed method, we choose the SU that
has the maximum Remaining Time (RT) after conducting
channel sensing, in order to balance the remaining idle
times for other tasks in the network.
For S1, the RT = (1000×10% - 50) = 50 ms, and for
S2, the RT = (1000×12% - 50) = 70 ms, and for S3, the
RT = (1000 × 15% - 100) = 50 ms. Therefore, the
maximum remaining time after sensing is 70 ms which
corresponds to SU neighbor S2.
Therefore, S2 is selected to do sensing for the licensed
channel, ch3 (which is used by SU4 to transmit data to
SU5), in behalf of SU4 when it is busy with data
transmission. It is worth mentioning that S3 is not
chosen, although S3 has the maximum idle time.
However, S2 is selected based on our selection metric
which is based on both factors: SU idle time and the
required channel sensing time.
V. System Model
In cognitive radio networks, SUs must perform in-
band sensing periodically in order to check whether the
currently used PUs channel becomes active again or not
[26]. This period of time is known in literature as
monitoring cycle.
This monitoring cycle frame, Tcyc, for a SU includes,
as shown in Fig. 4 and formulated in Eq. (1):
cyc tx sw sen busy idle
T T T T T T
    
(1)
- Data transmission time, Ttx, the available time of a
SU which can be used to transmit data.
- Switching time, Tsw, if needed (when the receiving
and transmitting channels are different).
- Sensing time, Tsen, in order to check whether the PU
becomes active again or not.
- Busy time, Tbusy, SU time utilized for other route
transmission and sensing, if exist.
- Remaining (idle) time, Tidle, SU is idle and this time is
not utilized.
In this work, our goal is to reduce the overall sensing
time for SUs over the selected route, in order to use this
saved time for data transmission instead.
Therefore, the route E2E delay is decreased and
throughput is increased, as explained in Section VI.
We assume SUs in the network have different sensing
time, even to the same channel.

S. H. Alnabelsi, R. R. Saifan, H. M. Almasaeid
Copyright © 2016 Praise Worthy Prize S.r.l. - All rights reserved International Review on Computers and Software, Vol. 11, N. 10
927
Fig. 4. A SU monitoring cycle time frame components
Also, SUs availability (idle) times are different, e.g.; if
a SU idle time is 30% of monitoring cycle time, this
means 30% of this SU time is not utilized, therefore, we
are motivated to use this idle time for cooperative
spectrum sensing.
Note that when SUs perform spectrum sensing, they
may employ energy detection technique which usually
needs about 1 ms, or employ feature detection technique,
e.g.; cyclostationary, which usually needs more than 1
ms, e.g.; 150 ms, [26], [27]. In feature detection
technique, SUs search for PUs features, e.g.; modulation
scheme employed by PUs that distinguishes them from
SUs transmission.
V.1. Reduction for the Overall Sensing Time Along
the Given Selected Route
In the proposed model, we reduce the overall sensing
time that performed by SUs along a selected route, if
possible, using sensing cooperation. We propose to select
for a SU along the selected route, a neighbor SU which
has some available (idle) time, in order to conduct
channel sensing. The neighbor SU required sensing time
must be less or equal to its idle time, therefore, this
neighbor SU can perform channel sensing.
The selection criteria of the cooperative SU is
formulated in Eq. (2), where a SU is selected from
neighbor SUs such that its remaining idle time, call it τ,
is the maximum after performing the cooperative
sensing. As Eq. (2) shows the available time for a
neighbor SU is calculated by the multiplication of the
monitoring cycle time and the SU neighbor idle time
percentage, which must be greater or equal to the
required sensing time by neighbor SU for the
corresponding PU channel:


i i i
cyc idle sen
max T T % T , i G

     (2)
V.2. Our Proposed Co-Operative Sensing Protocol
In our proposed cooperative sensing strategy, these are
the steps that are performed by each SU along the given
selected routing path, in order to select the neighbor SU,
if exist, to do a cooperative spectrum sensing.
Note that our proposed selection criteria is explained
in the previous subsection.
- Step 1: Initially, each SU on the given selected
routing path sends a control packet, call it “Hello”
packet, to its neighbor SUs to ask them about: (a) -
Their idle time percentage, (b) - Their required
sensing times for PUs channels within their
geographical area.
- Step 2: Neighbor SUs which receive the “Hello”
packet and within one-hop reply by the requested
information in step (1).
- Step 3: After that, each SU along the routing path
when receives the requested information, uses the
objective function formulated in equation (2), in order
to select a neighbor SU that does channel sensing on
its behalf.
Note that a neighbor SU, call it SUn, can perform a
cooperative sensing for a SU along the selected route,
call it SUr, if these conditions hold:
1. SUn geographical area is affected by the same PU(s)
transmission and sensing ranges as SUr.
2. SUn is affected by the exact set of PUs which has the
same licensed channel as SUr.
3. SUn available (idle) time is enough to perform
sensing for PU licensed channel.
Also note that a SUn may have no idle time, since it is
busy with other transmission or sensing for other
currently working routes in the network, called Tbusy.
After the cooperative SUs for spectrum sensing are
selected, at the end of each monitoring cycle, SUs along
the routing path communicate with their selected
cooperative neighbor SUs, in order to exchange
information about channels availabilities.
Exchanging control packets between a SUr and SUn is
typically less than the required sensing time when the SU
do spectrum sensing by itself. The exchange time that is
represented by Round Trip Time (RTT) is much less than
the required channel sensing time:
RTT = (2×packet transmission time) +
+ (2 × propagation delay) + transceiver
processing delay
(3)
such that:
- Packet transmission time = (packet size/channel
bandwidth),
- Propagation time = (distance/propagation speed),
- Propagation speed in wireless link = (3×108).
For example, say if the control packet size (which
contains information whether the PUs channels are busy
or not) = 60 bytes, bit rate = 1 Mbps, maximum distance
between SUs = 1000 meters, and the processing delay
(rounded within the SIFS) = 10 us [28]. By substituting
these values to find RTT, as follows:
6
6 8
2 60 8 2 1000
10 10 976.6us
1 10 3 10
RTT 
  
    
 
RTT calculation above shows that it is about (1 ms),
however, this time is still less than the required channel
sensing time by a SU which is usually more than (1 ms)

S. H. Alnabelsi, R. R. Saifan, H. M. Almasaeid
Copyright © 2016 Praise Worthy Prize S.r.l. - All rights reserved International Review on Computers and Software, Vol. 11, N. 10
928
and up to (150 ms) [26], [27], in order to check the PUs
activity over licensed channels.
VI. End-to-End Delay and Throughput
Enhancement
In this section, we discuss how our proposed method
reduces the E2E delay [30] for the selected route, since
the sensing time which is needed for SUs along the
selected route is reduced using the proposed cooperation
scheme. Eq. (4) is used to find the E2E delay of SUs for
a selected route:
1 1
2n n
i
x
i
i i
L
E E T
R U
 
 

  (4)
Let Ui denotes SUi utilization which represents the
remaining percentage of monitoring cycle time that SUi
can use for data transmission, Eq. (5).
Ui is increased by reducing sensing time. Therefore,
when our proposed cooperative sensing method is
employed, if possible, the sensing time becomes zero for
SUi such that in-band channel sensing is performed by
the selected cooperative neighbor SU:
i i i
cyc sen sw busy
icyc
T T T T
UT
  
 (5)
To find the bottleneck link that has the minimum rate
for the set of SUs along the route, use Eq. (6):


i
min U R , i g

   
(6)
when our proposed method is applied, the bottleneck
throughput, β, is enhanced or increased, because the Ui
for some SUs is most likely enhanced (increased).
Therefore, the E2E delay is decreased. Our simulation
results in Section VII show the effectiveness of our
proposed scheme.
VII. Simulation Results
In our simulation, PUs are distributed randomly in a
square area between (0, 0) and (4000, 4000), such that
the distances are in meters. SUs are distributed in a grid
based such that the square side length = 500 m, so the
total number of SUs is calculated as: (4000/500) *
(4000/500) = 64. The interference range for the PU is
650 m (as this notion is explained in Fig. 1, Section 3, the
transmission range of the SU is 500 m.
The simulation parameters are set as follows, unless
mentioned otherwise: channel’s sensing time by a SU is
selected randomly between 1 ms and 150 ms, the
message size delivered from the source SU to the
destination SU over the selected route is (1 Mb),
channel’s data rate = (10 Mbps), number of PUs = 30
distributed randomly in the area, number of licensed
channels in the network = 10, SU available time = 50%,
channel switching delay = 1 ms per 10MHz [29], the SU
monitoring cycle time = 1000 ms, probability of PU to be
active = 0.60.
This simulation evaluates three performance metrics
for the selected route:
1. The overall sensing time;
2. The E2E delay enhancement;
3. The improvement on the achievable bottleneck rate,
β.
Fig. 5 shows the enhancement percentages for the E2E
delay and the bottleneck, β, for different sensing time
distributions.
For example, when the sensing range for SUs is [0 -
130], the E2E delay and bottleneck enhancements are
6.6% and 11%, respectively.
Clearly in Fig. 5, when the sensing time range is
increased, the achieved enhancement is increased.
Fig. 6 shows the overall reduction percentages of the
required SUs sensing time along the route with different
SUs sensing time ranges.
Clearly, a novel improvement is achieved such that
the overall required sensing time along the route is
reduced by about 97% for SUs sensing ranges [0 - 40] to
[0 - 190], while it is 78% for [0 - 10] sensing range. Note
that sensing time for a channel by a SU ranges from 1 ms
to about 200 ms based on employed detection technique,
as explained in system model section V.
In Fig. 7, the E2E delay is reduced by about 7.5%.
Also, this figure shows the achieved bottleneck rate
improvement is increased from 9% (when number of
PUs = 10) to 11.3% (when number of PUs = 60), this
improvement demonstrates the necessity of employing
our proposed CSS, when number of PUs increases in the
network.
Fig. 8 shows an improvement for the route E2E delay
when CCS is employed by about 7.6% when the SUs
available time is randomly distributed between (0% and
40%) or more up to (0% and 90%).
Also in Fig. 8, it is shown the bottleneck throughput
for the route is improved by 1.2% when SUs available
time range [0 - 10], and improved by 12.5% when SUs
available time range [0 - 60].
Fig. 5. E2E delay and bottleneck rate enhancement
with different SUs sensing time ranges

S. H. Alnabelsi, R. R. Saifan, H. M. Almasaeid
Copyright © 2016 Praise Worthy Prize S.r.l. - All rights reserved International Review on Computers and Software, Vol. 11, N. 10
929
Fig. 6. Overall enhancement (reduction) percentages of the required
sensing time for SUs along the route, with respect to different SUs
sensing time ranges
Fig. 7. E2E delay and bottleneck rate enhancement with different
number of PUs
Fig. 8. E2E delay and bottleneck rate enhancement with different
SUs available percentages
Fig. 9. Overall enhancement (reduction) percentages of the required
sensing time for SUs along the route, with respect to SUs available
time percentage
Fig. 9 shows the overall reduction percentages of the
required sensing time for SUs along the route, with
respect to different SUs available time percentage.
A big improvement is achieved when SUs available
time range [0 - 30], the required overall sensing time is
reduced by 90%.
VIII. Conclusion
In Cognitive Radio Networks (CRNs), routing is
different from traditional network, because cognitive
radio users, also known as Secondary Users (SUs), are
required to have awareness about channels availability.
In this paper, we introduce a novel cooperative
spectrum sensing strategy, in which SUs along a selected
route cooperate with their neighboring SUs to monitor
Primary Users (PUs) activities. In the proposed model,
when a SU along a selected path wants to choose a
neighbor SU for cooperative channel sensing, SU
selection is based on: the required channel sensing time
(different for SUs), and the remaining available time of
the candidate SU. In our simulation, we study three
routing performance metrics: the overall sensing time,
the E2E delay enhancement, and the achieved
enhancement on bottleneck link rate, β.
Our results show that the proposed model reduces the
overall required channel sensing time along selected
routes, the bottleneck link rate is enhanced such that the
delivery rate is increased. Also, when SUs available time
increases, the performance metrics are enhanced.
Some challenges for our proposed protocol:
(a)-A SU along a path may not have SU neighbors, in
order to do cooperative sensing.
(b)-All neighbor SUs are busy and do not have enough
time to do cooperative sensing.
(c)-There is no SU that satisfies the required cooperation
condition, such that it must be affected by the same
set of PUs as SU along the path.
As a future work, the case when there is no neighbor
SU exists which is affected by the same set of PUs, we
search for more than one neighbor SU which are affected
by the same set of PUs, in order to improve sensing
performance.
References
[1] ET FCC, Docket no 03-222 notice of proposed rule making and
order, (2003).
[2] I. Akyildiz, W. Lee, M. Vuran, and S. Mohanty, Next
generation/dynamic spectrum access/cognitive radio wireless
networks: a survey, (2006) Computer Networks, Vol. 50, No. 13,
pp. 2127 - 2159.
[3] Mitola, J., Cognitive radio for flexible mobile multimedia
communications. In IEEE International Workshop on Mobile
Multimedia Communications (pp. 3 – 10, 1999)
[4] Ghasemi, A. and Sousa, E., Collaborative spectrum sensing for
opportunistic access in fading environments, IEEE International
Symposium on Dynamic Spectrum Access Networks (DySPAN),
(pp.131 – 136, 2005).
[5] X. Chen, Z. Bie, and W. Wu, Detection efficiency of cooperative
spectrum sensing in cognitive radio network, (2008) The Journal
of China Universities of Posts and Telecommunications, Vol. 15,

S. H. Alnabelsi, R. R. Saifan, H. M. Almasaeid
Copyright © 2016 Praise Worthy Prize S.r.l. - All rights reserved International Review on Computers and Software, Vol. 11, N. 10
930
Issue 3, pp. 1 - 7.
[6] Orumwense, E., Oyerinde, O., Mneney, S., Impact of Primary
User Emulation Attacks on Cognitive Radio Networks, (2014)
International Journal on Communications Antenna and
Propagation (IRECAP), 4 (1), pp. 19-26.
[7] Anusha, M., Srikanth, V., An Efficient Mac Protocol for
Reducing Channel Interference and Access Delay in Cognitive
Radio Wireless Mesh Networks, (2016) International Journal on
Communications Antenna and Propagation (IRECAP), 6 (1), pp.
14-18.
[8] Belghiti, I., Elmachkour, M., Berrada, I., Omari, L., Green
Cognitive Radio Networks by Using Coalitional Game Approach
in Partition Form, (2014) International Review on Computers and
Software (IRECOS), 9 (10), pp. 1705-1711.
[9] Sebgui, M., Bah, S., Elgraini, B., A Prediction-Based Solution for
Improving the Performance of CSMA/CA Networks Under the
Hidden Collision Effect, (2015) International Review on
Computers and Software (IRECOS), 10 (7), pp. 710-717.
[10] Alnabelsi, S.H. and Kamal, A.E., Interference-based packet
recovery for energy saving in Cognitive Radio Networks, IEEE
ICC workshop on Cognitive Radio and Cooperation for Green
Networking (pp. 5978-5982, 2012).
[11] Alnabelsi, S.H. and Kamal, A.E., Resilient multicast routing in
CRNs using a multilayer hyper-graph approach, IEEE
international conference on communications (pp. 2910-2915,
2013).
[12] Alnabelsi, S.H., Kamal, A.E. and Jawadwala, T.H., Uplink
channel assignment in Cognitive Radio WMNs using physical
layer network coding, IEEE International Conference on
Communications (pp. 1-5, 2011).
[13] S. Althunibat, M.D. Renzo, and F. GranelliM, Cooperative
spectrum sensing for cognitive radio networks under limited time
constraints, (2014) Computer Communications, vol. 43, pp. 55 -
63.
[14] S. Atapattu, C. Tellambura, and H. Jiang, Energy Detection Based
Cooperative Spectrum Sensing in Cognitive Radio Networks,
(2011) IEEE Transactions on Wireless Communications, Vol. 10,
No. 4.
[15] Z. Quan, S. Cui, and A. Sayed, Optimal Linear Cooperation for
Spectrum Sensing in Cognitive Radio Networks, (2008) IEEE
Journal of Selected Topics in Signal Processing, Vol. 2, No. 1.
[16] Sun, C., Zhang, W., and Ben Letaief, K., Clusterbased
cooperative spectrum sensing in cognitive radio systems, IEEE
International Conference on Communications (ICC), (pp. 2511-
2515, 2007).
[17] S. Maleki, A. Pandharipande, and G. Leus, Energy-Efficient
Distributed Spectrum Sensing for Cognitive Sensor Networks,
(2011) IEEE Sensors Journal, Vol. 11, No. 3, pp.565 -574.
[18] R. Deng, J. Chen, C. Yuen, P. Cheng, and Y. Sun, Energy-
Efficient Cooperative Spectrum Sensing by Optimal Scheduling
in Sensor-Aided Cognitive Radio Networks, (2012) IEEE
Transactions On Vehicular Technology, Vol. 61, Issue 2, pp. 716
- 725.
[19] S. Eryigit, S. Bayhan, and T. Tugcu, Energy-Efficient
Multichannel Cooperative Sensing Scheduling With
Heterogeneous Channel Conditions for Cognitive Radio
Networks, (2013) IEEE Transactions of Vehicular Technology,
Vol. 62, No. 6.
[20] Su, H. and Zhang, X., Energy-Efficient Spectrum Sensing for
Cognitive Radio Networks, IEEE International Conference on
Communications (ICC), (pp. 1-5, 2010).
[21] N. Zhao, H. Sun, and A. Nallanathan, Energy efficient
cooperative spectrum sensing schemes for cognitive radio
networks, (2013) EURASIP Journal on Wireless Communications
and Networking.
[22] S. Li, Z. Zheng, E. Ekici, and N. Shroff, Maximizing System
Throughput by Cooperative Sensing in Cognitive Radio
Networks, (2011) IEEE/ACM Transactions on Networking, vol.
22, no. 4, pp. 1245 - 1256.
[23] V. Rakovic, D. Denkovski, V. Atanasovski, P. Mhnen, and L.
Gavrilovska, Capacity-Aware Cooperative Spectrum Sensing
Based on Noise Power Estimation, (2015) IEEE Transactions on
Communications, vol. 63, no. 7, pp. 2428 - 2441.
[24] E. Peh, Y. Liang, Y. Guan, and Y. Zeng, Optimization of
Cooperative Sensing in Cognitive Radio Networks: A Sensing-
Throughput Tradeoff View, (2009) IEEE Transactions on
Vehicular Technology, Vol. 58, No. 9.
[25] Saifan, R., Kamal, A. E., and Guan, Y., Efficient Spectrum
Searching and Monitoring in Cognitive Radio Network, In the
proceedings of the IEEE International Conference on Mobile Ad-
hoc and Sensor Systems (MASS), (2011).
[26] Kim, H. and Shin, K., In-band Spectrum Sensing in Cognitive
Radio Networks: Energy Detection or Feature Detection?, The
14th Annual International Conference on Mobile Computing and
Networking, (pp. 14 – 25, 2008).
[27] Pandya, P., Durvesh, A. and Parekh, N., Energy Detection Based
Spectrum Sensing for Cognitive Radio Network, Fifth
International Conference on Communication Systems and
Network Technologies (CSNT), (pp. 201 – 206, 2015).
[28] IEEE Standard 802.11, Part 11: Wireless LAN Medium Access
Control (MAC) and Physical Layer (PHY) Specifications, (2012).
[29] TCI International (2010). “Tci 715 spectrum monitoring system
data specification”.
http://www.tcibr.com/ufiles/File/715Webp.pdf.
[30] Ahmed, F., El Mashade, M., Konber, H., Quality of Service
Enhancement of Vice Over Internet Protocol Via Novel
Scheduling Algorithm, (2015) International Journal on
Communications Antenna and Propagation (IRECAP), 5 (6), pp.
375-378.
Authors’ information
1Al-Balqa’ Applied University.
2Jordan University.
3Yarmouk University.
Dr. Sharhabeel Alnabelsi is an assistant
professor at Computer and Networks
Engineering Dept. at Al-Balqa' Applied
University, Jordan. He received his Ph.D. in
computer engineering from Iowa State
University, USA in 2012. Also, he received his
M.Sc. in computer engineering from The
University of Alabama in Huntsville, USA in
2007. His research interests include Cognitive Radio Networks,
Wireless Sensor Networks, and network optimization.
Dr. Ramzi Saifan received his B.Sc. and M.S.
degrees in computer engineering from Jordan
University of Science and Technology, Irbid,
Jordan, in 2003 and 2006 respectively, and
received his Ph.D. degree in computer
engineering from Iowa State University in 2012,
USA. Since Jan 2013, he joined as an assistant
professor in the Department of Computer
Engineering at the University of Jordan, Jordan. His current research
interests include computer networks, computer and network security,
cognitive radio networks, and image processing. Saifan published
several papers in peer reviewed journals and conferences
Dr. Hisham M. Almasaeid is an assistant
professor of Computer Engineering at Yarmouk
University, Jordan. He received his Ph.D. from
Iowa State University in Fall 2011, USA. His
research interests include Cognitive Radio
Networks, Wireless Sensor Networks, and
Mobile Ad Hoc Networks.