TRIBHUVAN UNIVERSITY
INSTITUTE OF ENGINEERING
PULCHOWK CAMPUS
THESIS NO.: 069/MSICE/610
DYNAMIC SPECTRUM CO-ACCESS (DSCA) WITH DIRTY PAPER
CODING (DPC) FOR COGNITIVE RADIO NETWORK
BY
NASHIB ACHARYA
A THESIS
SUBMITTED TO THE DEPARTMENT OF ELECTRONICS AND COMPUTER
ENGINEERING IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE
DEGREE
OF
MASTER
OF
SCIENCE
IN
INFORMATION
AND
COMMUNICATION ENGINEERING
DEPARTMENT OF ELECTRONICS AND COMPUTER ENGINEERING
November 2014
DYNAMIC SPECTRUM CO-ACCESS (DSCA) WITH DIRTY
PAPER CODING (DPC) FOR COGNITIVE RADIO NETWORK
Submitted By:
Nashib Acharya
(069/MSICE/610)
Thesis Supervisor
Dr. Nanda Bikram Adhikari
A thesis submitted in partial fulfillment of the requirements for
the degree of Master of Science in Information and
Communication Engineering
Department of Electronics and Computer Engineering
Institute of Engineering, Pulchowk Campus
Tribhuvan University
Lalitpur, Nepal
November 2014
ACKNOWLEDGEMENT
I am very much thankful to the Department of Electronics and Computer
Engineering, IOE, Pulchowk campus for providing the opportunity for presenting
myself with this Thesis entitled “Dynamic Spectrum Co-Access (DSCA) With Dirty
Paper Coding (DPC) For Cognitive Radio Network”
I express my deepest appreciation to my thesis supervisor Dr. Nanda Bikram
Adhikari, who has the attitude and the substance of a genius. I am very thankful for
his expert guidance, suggestions and coordination. Without his guidance and
persistent help this dissertation would not have been possible.
I would like to express special thanks of gratitude to the Head of Department of
Electronics and Computer Engineering, IOE, Pulchowk Campus, Assistant Professor
Dr. Dibakar Raj Pant, for his guidance and support.
I am very grateful to our Program Coordinator of Masters of Science in Electronics
and Communication Engineering, Assistant Professor, Mr. Surendra Shrestha for his
kind support and valuable coordination.
It would be injustice without acknowledging our Former Program Coordinator of
Masters of Science in Electronics and Communication Engineering and Vice Campus
Chief,Assistant Professor, Mr. Sharad Kumar Ghimire for his precious coordination.
I would like to express my gratitude to all the Professors and Teachers of Institute of
Engineering (IOE) who have helped and guided me directly or indirectly.
Finally, I would like to thank all the faculty member of Kathford International
College of Engineering and Management, friends and colleagues of MSICE for their
precious encouragement and support.
i
ABSTRACT
In the current architecture of dynamic spectrum access, which is also known as
opportunistic spectrum access, secondary users only opportunistically access the
spectrum of primary users. The resurgence of primary users disrupts on going
communication of secondary users, which can result in poor performance for
secondary users. In this thesis, an architecture for dynamic spectrum access, termed
Dynamic Spectrum Co-Access (DSCA), is implemented to enable the primary user
and the secondary user to simultaneously access licensed spectrum. With DSCA,
secondary users transparently incentivize primary users through increasing the
primary user performance, so that secondary users can access spectrum
simultaneously with primary users; hence there is no disruption to secondary
communications due to the resurgence of primary users. A mathematical model is
formulated to calculate the minimum incentives for the spectrum co-access between
the primary user and the secondary user, and the region of co-access is computed to
determine the secondary users that can co-access with a given primary user. 900 Mhz
band is used for the study purpose, especially for the power allocation of the
secondary user. Special pre coding techniques called Dirty Paper Coding (DPC) is
used to preserve signal over the interference for secondary network. A mathematical
model is formulated to determine the minimum incentives for the spectrum co-access,
computational analysis of region of co-access is done to determine where the
secondary users can co-access with a given primary user.
Keywords: Dynamic spectrum Co-Access; simultaneous access of spectrum;
cognitive radio network; resurgence; disruption.
ii
RECOMMENDATION
The undersigned certify that they have read and recommended to the Department of
Electronics and Computer Engineering for acceptance, a thesis entitled “Dynamic
Spectrum Co-Access (DSCA) With Dirty Paper Coding (DPC) For Cognitive
Radio Network”, submitted by Nashib Acharya in partial fulfillment of the
requirement for the award of the degree of “Master of Science in Information and
Communication Engineering”.
……………………………………
Supervisor:
Dr. Nanda Bikram Adhikari
Assistant Professor
Department of Electronics and Computer Engineering
Institute of Engineering
Pulchowk campus
iii
TABLE OF CONTENTS
ACKNOWLEDGEMENT ……………………………………….…………………….i
ABSTRACT …………………………………………………………………………..ii
RECOMMENDATION ………………………………………………………………iii
TABLE OF CONTENTS……………………………………………………………..iv
LIST OF FIGURES ...…………………………………………………………………v
LIST OF TABLES …………………..……………………………………………….vii
ABBREVIATIONS …………………………………………………………………viii
CHAPTER 1: INTRODUCTION .…………………………………………………….1
1.1 Background ……………………………………………………………………2
1.2 Problem Statement .……………………………………………………………4
1.3 Objectives ……………………………………………………………………..5
CHAPTER 2: LITERATURE REVIEW ……………………………………………...6
CHAPTER 3: RELATED THEORY ………………………………………………....9
3.1 Existing Cognitive Network Paradigm ………………………………………10
3.1.1 Interweave paradigm ………………………………………………….10
3.1.2 Underlay paradigm ……………………………………………………11
3.1.3 Overlay paradigm ……………………………………………………..11
3.2 Interference Channel: An Overview …………………………………………12
3.3 Dynamic Spectrum Co-Access ......…………………………………………..13
3.4 Co-Access Encoding Techniques …………………………………………....14
CHAPTER 4: RESEARCH METHODOLOGY…………………………………….15
iv
4.1 One PU Node Pair and One SU Node Pair ………………………….…...….16
4.1.1 At PU transmitter ………………………………………………….….18
4.1.2 At SU transmitter ………………………………………………….….18
4.1.3 At PU receiver ………………………………………………………..19
4.1.4 At SU receiver ………………………………………………………..20
4.2 Co Access Incentive ………………………………………………………....21
4.3 Acceptable SU SINR ………………………………………………………...21
4.4 Region of Co-Access ………………………………………………………...22
CHAPTER 5: SIMULATION RESULTS AND INTERPRETATION……………..24
5.1 Simulation Parameters ……………………………………………………...25
5.2 Achievable Rates …………………………………………………………….26
5.3 Region of Co-Access …………………………………………………….…..30
CHAPTER 6: EPILOGUE …...………………………………………………………32
6.1 Limitations …………………………………………………………………...33
6.2 Future Enhancements ………………………………………………………...33
6.3 Schedule ……………………………………………………………………...33
6.4 Conclusions …………………………………………………………………..34
CHAPTER 7: REFERENCES ………………………………………………………35
v
LIST OF FIGURES
Figure 1.1 Basic Cognitive Cycle …………………………………………………….3
Figure 3.1 Conceptual illustration to show interference channel…………………….13
Figure 3.2 An incentivized co-access architecture…………………………………...14
Figure 4.1 Basic Incentivized Architecture with a PU and a SU node pair. …………16
Figure 4.2 Architecture to determine region of co-access consisting three PU node
and a SU node ………..…………………………………………………..23
Figure 5.1 Plot of maximum achievable PU rate with increasing 𝛾 while PU
transmitting 5 watt and SU transmitting 7 watt. …………………......…..26
Figure 5. 2 Plot of maximum achievable SU rate with increasing 𝛾 while PU
transmitting 5 watt and SU transmitting 7 watt. …………......…………..27
Figure 5.3 Figure showing the comparative change in maximum achievable rate of
PU and SU with change in 𝛾 ………………………..........………………28
Figure 5.4 Effect of change in transmit power level of SU on maximum achievable
rate of SU network ………………………………..........………………...29
Figure 5.5 Effect of change in transmit power level of SU on maximum achievable
rate of PU network ……………………......……………………………...30
Figure 5.6: Region of co-access where SU can located between any two PU node …31
vi
LIST OF TABLES
Table 4.1: Major notations used to formulate the mathematical model……………...18
Table 5.1: Table showing values of transmit power level of base station in dBm and
watt for GSM 900 ………………………………..………………………25
Table 6.1: Time schedule …………………………………………………………….33
vii
LISTS OF ABBREVIATIONS
CRN
Cognitive Radio Network
DPC
Dirty Paper Coding
DSA
Dynamic Spectrum Access
DSCA
Dynamic Spectrum Co-Access
FCC
Federal Communication Commission
GSM
Global System of Mobile
PU
Primary User
QoS
Quality of Service
SU
Secondary User
SINR
Signal to interference Plus Noise Ratio
SNR
Signal to Noise Power Ratio
TV
Television
UHF
Ultra High Frequency
VHF
Very High Frequency
viii
CHAPTER ONE
INTRODUCTION
1
1.1 Background
Broadcasting is done through Radio. Thus, large number of users coexists in same
frequency band which interfere each other. As numbers of users increased
exponentially over last few decades, the availability of spectrum becomes severely
constraint. This shows that almost all the frequency bands have been assigned. On the
other hand, user demands are increasing exponentially. Thus, it is obvious that the
availability of the spectrum becomes severely constraint.
In recent years, significant effort has been applied to better utilize the wireless
communications spectrum. The existing model for spectrum allocation by FCC has
been to give licenses for the major part of usable spectrum to the commercial licensed
user and named them as primary user. The purpose of these unlicensed bands is to
encourage innovation without the high cost to entry associated with purchasing
licensed spectrum through auctions. The unlicensed bands have been proven a great
vehicle for innovation, and the 2.4 GHz unlicensed band currently host systems such
as Bluetooth, 802.11b/g/n Wi-Fi, and cordless phones. Unfortunately, the unlicensed
bands can be killed by their own success, since the more devices that occupy these
bands, the more interference they cause to each other. As PU pay for the spectrum,
the total right over this spectrum will be of PU. However, many studies have shown
that a large portion of licensed spectrum is underutilized. There exist abundant
spectrum opportunities in the temporal, spatial, and frequency domains. The
exploitation of these spectrum opportunities is currently an area of significant
research known as DSA or CRN. Researchers consider cognitive radio as the best
solution for the problem of spectrum scarcity, since a large portion of spectrum in the
UHF/VHF bands are becoming available on a geographical basis after analog to
digital TV switchover [1].
There exist very little new bandwidth available for
emerging wireless products and services. Cognitive radio is born as the idea to solve
this spectrum scarcity problem. As stated by Andrea Goldsmith, Syed Ali Jafar and
2
Ivana Maric, “A cognitive radio is a wireless communication system that intelligently
utilizes any available side information about the (a) activity, (b) channel conditions,
(c) codebooks or (d) messages of other nodes with which it shares the spectrum” .
This is the most famous and widely proposed cognitive radio cycle. However, it is the
concept which is basically used in opportunistic cognition. The CR devices will
utilize advanced radio and signal processing technology along with new spectrum
allocation policies to support new users in existing crowded spectrum without
degrading the QoS of the existing users of the spectrum. A CR must pose advanced
sensing and processing capabilities Thus CR needs the intelligence software which
can sense, gather and process all the information about the spectrum that exists
around it [2]. Thus CR is a concept that allows wireless system to sense the
environment, adapt, and learn from previous experience to improve the
communication quality.
Figure 1.1 Basic Cognitive Cycle
In opportunistic spectrum access, SU opportunistically access the licensed
spectrum of PU, whereas PUs has privileged access of the licensed band. The
3
compulsion to vacate the band immediately by SU after resurgence of primary user
traffic in the band made ongoing communication of secondary user to be disrupted.
The requirement that Secondary users cannot access spectrum simultaneously with
Primary users results in significant overhead on spectrum sensing and spectrum
handoff, which in turn results poor performance for cognitive radio networks.
In my Thesis, A novel architecture is developed for dynamic spectrum access, called
DSCA, which enables SU simultaneously access licensed spectrum with PU through
transparently incentivizing PU network. It differs than opportunistic spectrum access
in a way that it allows simultaneous access not time based sharing. DPC is
incorporated as a precoding technique with cooperative cognitive radio to implement
this. It is well understood that PUs are not willing to share their licensed spectrum
without incentives. The novelty of DSCA is that the secondary user communication
can provide a significant performance improvement to the PUs communication as
incentive. Hence PU is incentivized to welcome the co-access of spectrum with SUs
[3].
1.2 Problem Statement
Cognitive radio was first born to solve the spectrum scarcity problem, however in
existing opportunistic spectrum access as the traffic of primary user reappear in the
licensed band, it compels secondary transmitter to terminate its ongoing
communication which result very poor quality of service of cognitive network as the
resurgence of PU is very obvious and frequent. Following are some of the problem
statements that led situation for dissertation to develop.
1. Higher data rate demand of wireless users.
2. The resurgence of primary user in traffic band disrupts ongoing
communication of secondary users.
4
3. Primary users are not willing to share their spectrum without any
incentives.
4. Lack of simultaneous access of spectrum in opportunistic spectrum access
architecture.
1.3 Objectives
In contrast to opportunistic spectrum access, this thesis aims to implement
simultaneous access of PU and SU to the spectrum. The main objectives of this thesis
that enables the simultaneous access of spectrum are as follows:
1. To implement incentivized Spectrum access among Primary user and
Secondary User.
2. To calculate the achievable rate of both PU and SU while SU Co-Access
with PU.
3. To determine the Region of Co-Access where SU can Co-Access with PU.
5
CHAPTER TWO
LITERATURE REVIEW
6
The concept of cognitive radio was first proposed in 1998 in the seminar of Royal
Institute of Technology of Stockholm. It describes that if the network is intelligent
enough to gather the information about the co-users, then the radio resources can be
adaptively changed to meet users need and demand. The cognitive radio is defined as
the goal towards software defined radio platform [4]. A lot of researchers have stated
in their article that the available white space of the TV band could be increased to 120
MHZ for cognitive users after the analog to digital TV switchover [5]. This shows the
possibility of cognitive radio. In the past there have been extensive studies on
opportunistic spectrum access architecture and cognitive radio networks. Good
general overview of the cognitive radio can be found in [6] and [7]. Different
paradigm of the cognitive network is briefly discussed and concluded various aspects
of the paradigm which are already mentioned in introduction section. Underlay and
Overlay allows concurrent transmission of both primary and secondary user in Along
with it, this paper also explain various encoding techniques and error control
techniques for the interference cancellation during concurrent transmission. This
suggests dirty paper coding as the best candidate for encoding in the cognitive radio
network in known interference scenario [2]. The network coding technique is
discussed and implemented to incentivize PU to cooperate with SU in spectrum
access, so that SUs can access spectrum even when PU is active. Nevertheless, the
spectrum access of SUs is not transparent to PUs in this scheme. The PUs must have
the knowledge of SUs, and need to listen to the packets from SU [3]. Contrary to the
scheme in [3], the spectrum access of SU in DSCA architecture in this thesis is
transparent to PUs, i.e. PU does not need to have any knowledge of SU. The DSCA
architecture utilizes the DPC technique to achieve transparent incentivizing of PU.
DPC was first introduced by Costa as a proof for maintaining SINR at the receiver
given the transmitter had prior knowledge of the interference state. It was shown that
DPC could achieve the largest known capacity region for cognitive radio networks in
a channel model with one PU node pair and one SU node pair, as long as the SU
transmitter had a priori knowledge of the PU messages . Several later studies have
7
shown that SUs can coexist with PUs without degrading the PU channel capacity
.However the success of DPC in a cognitive radio network relies on the SU
transmitter having a priori knowledge of the PU transmitted packet [8]. This is a nontrivial problem and there have been several proposed methods for achieving this. In
traditional one-hop infrastructure networks the authors proposed using DPC for
interference reduction between base stations, by leveraging the high bandwidth of the
wired backbone to obtain a priori knowledge of base stations downlink data. However
the PU is unlikely to share a wired high-bandwidth backbone with SUs [9].
More importantly, the Dirty Paper Coding have been implemented to cancel the joint
interference for the cognitive users. In their architecture primary codebooks are
known to the secondary base station and interference cause by the primary base
station on the secondary users is canceled using DPC [10].
In this thesis, DSCA for transparent incentivizing of primary user by secondary user
has been implemented. PU need not to be aware of existence of secondary user.
Moreover, secondary user operate themselves in such a way that it improves the
performance of primary user instead of interfering them. This now enables Primary
and secondary user to have simultaneous access of the spectrum which significantly
reduces the sensing and handoff overhead ensuring the QoS of cognitive radio
network.
8
CHAPTER THREE
RELATED THEORY
9
3.1 Existing Cognitive Networks Paradigms
There are three main cognitive radio network paradigms: Underlay, Overlay, and
Interweave. In interweave paradigm the cognitive radios opportunistically exploit
spectral holes to communicate without disrupting other transmission. In contrast to
that, the underlay paradigm allows concurrent transmission of primary and secondary
users. In underlay, cognitive users are allowed to operate if the interference by them
caused to non-cognitive users is below a given threshold i.e. interference temperature
is acceptable for primary users. In overlay systems the cognitive users use
sophisticated signal processing and coding to maintain or improve the SINR of noncognitive radios while also obtaining some additional bandwidth for their own
communications. Underlay and overlay are sometimes also called cooperative
cognitive radios.
3.1.1 Interweave paradigm
The interweave Paradigm is based on the principle of opportunistic communication,
which was the basic motivation of the cognitive radio. This was born after the study
conducted by FCC [11] and industry [12] which showed that the major part of the
spectrum is not utilized most of the time. This temporarily exist space – time –
frequency voids, referred as spectrum holes which are sensed and communication is
performed using these spectrum holes. This spectrum holes are further conditionally
termed as white space. Generally 1 MHz bandwidth available for ten minute, then it is
called white space. The nature of spectrum holes will change according to the time
and geographic location and thus can be exploited by cognitive users which finally
increase the spectrum utilization. This technique requires the knowledge of the side
activity of the non-cognitive users. Thus interweave technique can be defined as the
intelligence system that detects the occupancy of the spectrum periodically and
utilizes the spectrum holes with minimal acceptable interference to the active users or
10
primary users. The periodic sensing should be done due to changing nature of the
spectrum holes.
3.1.2 Underlay paradigm
In an underlay paradigm, the major motivation of this paradigm is that it allows
simultaneous cognitive and non-cognitive communication under the constraints that
the cognitive users are assumed to have the knowledge of the interference caused by
the cognitive transmitter to the non-cognitive user. So in this paradigm the cognitive
user (often called secondary user) cannot significantly interfere with the existing
communication device (often called primary user). Being specific, the concurrent
non-cognitive and cognitive transmission is allowed only if the interference level
caused by secondary user is below the acceptable threshold for the primary user.
3.1.3 Overlay paradigm
Unlike interweave paradigm, it also allows simultaneous transmission in the same
frequency band. In an overlay paradigm, the cognitive transmitter is assumed to have
knowledge of the non-cognitive users’ codebooks and its messages as well. This is
possible if the uniform standard is followed by the non-cognitive user or they could
transmit their codebooks periodically. Knowledge of a non-cognitive user’s
codebooks and messages can be exploited in various ways to either cancel or mitigate
the interference seen at the cognitive and non-cognitive users. On the other part, this
information can be used to completely cancel the interference due to non-cognitive
signals at the cognitive receiver by various techniques [2]. The cognitive users can
utilize the knowledge they have about the non-cognitive communication and use part
of their power for their own communication and the remainder of the power to relay
the non-cognitive transmission. By careful choice of power split, the increase in the
signal to noise ratio power ratio of primary user due to the additional signal from the
11
cognitive radio relaying. This increased SINR will guarantees that the interference
level created by remaining of the transmit power of cognitive users to non-cognitive
receiver does not change the non-cognitive user’s rate. This paradigm can be used in
both licensed and unlicensed band. In licensed bands, Cognitive users would be
allowed to share the band with licensed users since they do not interfere or even
might improve their communication. In unlicensed bands cognitive users would
enable a higher spectral efficiency by exploiting message and codebook knowledge to
reduce interference.
Overlay cognitive radio networks allow concurrent cognitive and non-cognitive
transmissions, also in contrast to underlay networks; the cognitive transmitter may
now facilitate the transmission of non-cognitive user also. Thus, smallest overlay
cognitive radio network is a two user (cognitive and non-cognitive) interference
channel where the cognitive transmitter has non causal knowledge of the noncognitive user’s messages. Knowledge of the non-cognitive user’s message allows the
cognitive transmitter to apply several encoding schemes that will improve both its
own rates as well as the rate of non-cognitive users. Various forms of coordination
are possible [2].
3.2 Interference Channel: An Overview
The interference channel model captures scenarios in which multiple terminal pairs
wish to communicate simultaneously in the presence of mutual interference. The
users are not assumed to be cognitive - they do not monitor the activity or decode
messages of other users. However, it is commonly assumed that all terminals know
the channel gains and the codebooks of all the encoders. The communication problem
is to determine the highest rates that can simultaneously be achieved with arbitrarily
small error probability at the desired receivers, i.e., to determine the capacity region.
This performance can serve as a benchmark to evaluate the gains of cognition. Even
12
for the smallest interference network consisting of two transmitter-receiver pairs, this
problem has remained unsolved for more than thirty years, emphasizing that one of
the fundamental problems in networks – coping with and exploiting interference - is
not yet entirely understood. Still, there has been a lot of progress in understanding
communication in interference channels [2].
Message W1
T1
Code word X(W1)
Estimate of W1
R1
Message W2
Estimate of W2
T2
R2
Code Word X(W1)
Figure 3.1: Conceptual illustration to show interference channel
3.3 Dynamic Spectrum Co- Access
In DSCA, when PU is not transmitting, SU freely access the spectrum as in
opportunistic spectrum access. But, while PU is active, SU provide incentives to
Primary for having co-access to the spectrum. Region of interest is of latter case, i.e.
how the SU incentivized the Primary user to enable spectrum co-access. The two
main points are co-access incentives and region of co-access. Incentives will be in the
form of increased performance of primary user. This is done by secondary user by
using part of its power to relay the message of the primary user and rest to transmit its
own message. The co-access incentives ensure that both PU and SU benefit from coaccess. The region of co-access defines where the SU can co-access with PU.
13
Message W1
T1
Code word X (W1)
Estimate of W1
R1
Message W2
Estimate of W2
T2
R2
Code X (W1, W2)
Figure 3.2: An incentivized co-access architecture
3.4 Co-Access Encoding Techniques
Co-access encoding techniques have mostly been investigated for the interference
channel with one cognitive encoder, which is also known as basic overlay network.
The special case of this co-access model is the basic interference channel i.e. when
both user is non-cognitive. An elegant idea of superposition coding [13] allows
sending information simultaneously to all users at higher data rates than what can be
achieved with time sharing. The DSCA architecture utilizes the DPC technique to
achieve transparent incentivizing of PUs.
DPC was first introduced by costa as a proof for maintaining SINR at the receiver
given that transmitter had prior knowledge of the interference state [8]. DPC allows
encoder to pre-code its message at a rate associated with interference-free
communication. It is called dirty paper because it is similar to writing in a paper with
dirt already (coding in the channel with existing interference).
14
CHAPTER FOUR
RESEARCH METHODOLOGY
15
Dynamic spectrum co-access architecture is developed to enable simultaneous
transmission of primary and secondary user. Matlab is used for the simulation
purpose.
In this section, DSCA architecture is described. With DSCA when PU is not
transmitting, SU freely access the spectrum, similarly to the opportunistic spectrum
access architecture. On the other hand, when PU is active, SU provide incentives to
PU so that simultaneous transmission by SU is allowed. In the following, operation of
DSCA in the latter case is focused, i.e., how the SU incentivizes the PU to enable
spectrum co-access. The SU allocates the portion of its power to PU and remaining
for transmitting its own code word. So PU receiver receives the code word from both
PU transmitter as well as SU Transmitter. The power allocation should be done in
such a way that PU’s SINR should be increased which is termed to be incentivized.
At first a simple network with one PU node pair and one SU node pair is considered.
Three key components of DSCA is used, Portion of SU power used to relay the PU
message, co-access incentives and region of co-access. The co-access incentives
ensure that both PUs and SUs are benefit from the spectrum co-access. The region of
PUtx
1
PUtx
b
b
a
SUtx
2
SUtx
Figure 4.1: Basic Incentivized Architecture with a
PU and a SU node pair
co-access is the region where SU can be located to co-access spectrum with PU.
Figure 4.1 show the basic architecture of an incentivized network with one SU node
16
and one PU node with normalized (1, a, b, 1) channel. The legend on a link indicates
the path loss.
4.1 One PU Node Pair And One SU Node Pair
In above given basic architecture of incentivized network, let Xp and Xs be the
codeword transmitted by PU and SU respectively. (1, a, b, 1) is assumed as a
normalized path loss between the links. Assume that SU knows the PU packet priori
through a side information path. To provide incentives to the PU, so that the PU
allows simultaneous spectrum access for SU, the SU transmitter uses a portion of its
power to boost the SINR at the PU receiver. Let γ ∈ [0, 1] denote the portion of the
SU power used to transmit the PU code word and (1−γ) the portion of power used to
transmit its own code word. Let Pp and Ps denote the transmit power of the PU and
SU transmitters, respectively. In addition, let Xp and Xs be a single transmitted code
word for the PU and SU, respectively. The major notations are listed in Table 4.1.
Over a large set of code words, the PU transmit power at the PU transmitter is PP =
|Xp|2. As it is already mentioned that SU have priori knowledge of PU, the DPC
technique is used to form the final codeword That emerge from the SU transmitter
which consist codeword of both SU and PU. The SU code word is generated using
DPC such that:
̃𝑠 + 𝑋𝑝 √𝛾𝑃𝑠 ,
𝑋𝑠 = 𝑋
𝑃
(4.1)
𝑝
̃𝑠 is the code word to carry the SU packet and 𝑋𝑝 √𝛾𝑃𝑠 is the code
where 𝑋
𝑃
𝑝
word to carry the PU packet. For the large set of code word, these code words are
chosen in such a way that they are statically independent, if they are not statistically
independent then they SU receiver will fail to cancel the interference of the PU. A
random bining technique is used for the choosing of these codeword. Table 4.1 shows
major notations summary used in this chapter.
17
a,b
γ
Normalized path losses as shown in figure 4.1
Portion of the SU power used to relay the PU code word
Pp,Ps
Transmit Power of the PU and SU transmitter respectively
𝑆𝑃, 𝑆𝑆
𝑄𝑝 ,𝑄𝑠
𝑋𝑝 , 𝑋𝑠
̃𝑠
𝑋
Received codeword by PU and SU receiver respectively
Received signal power (excluding interference) at the PU and SU
receivers, respectively
Transmitted code word of PU and SU transmitters
Code word of SU transmitter to carry SU packet
𝑅𝑝 , 𝑅𝑠
Achievable rate of PU and SU respectively
Np, Ns
Noise plus interference Power at PU and SU
Table 4.1: Major notations used to formulate the mathematical model
4.1.1 At PU transmitter
As stated earlier, this DSCA architecture transparently incentivized the PU i.e.
PU don’t need to be aware of the existence of SU, so no difference will be seen in the
nature of PU transmitter then it was without the SU. So the transmit power of the
Primary User can simply be given as:
PP = |Xp|2
(4.2)
4.1.2 At SU transmitter
The codeword transmitted by the SU transmitter consist the two code word
separately. One its own code word and another to relay the codeword of Primary User
for the incentive. This is given by the equation 4.1. So total power transmitted by the
SU transmitter is:
18
2
γP
̃𝑠 + 𝑋𝑝 √ s ) ,
Ps = (𝑋
𝑃
(4.3)
𝑝
̃𝑠 ⌉2 + 2𝑋
̃𝑠 𝑋𝑝 √γPs + 𝑋𝑝 2 . γ𝑃𝑠 ,
Ps =⌈𝑋
𝑃
𝑃
𝑝
̃𝑠 ⌉2 + γ𝑃𝑠 ,
Ps =⌈𝑋
𝑝
(4.5)
(4.4)
̃s X p = 0, they are statically independent]
[X
̃𝑠 ⌉2 = (1- γ) 𝑃𝑠 .
⌈𝑋
(4.6)
4.1.3 At PU receiver
The received signal at PU receiver will be the sum of signal transmitted by PU
and the sum of code word transmitted by SU. Received code word will be:
̃𝑠 + 𝑋𝑝 √γPs ),
𝑆𝑝 = 𝑋𝑝 + 𝑎 (𝑋
𝑃𝑝
γP
̃𝑠 .
𝑆𝑝 = (𝑋𝑝 + 𝑎𝑋𝑝 √ 𝑃 s ) + 𝑎𝑋
𝑝
Desired Code
(4.7)
(4.8)
Noise
The total desired signal power can be calculated from the Eq. (4.8) by
squaring the desired code word and is given by:
𝛾𝑃
𝑄𝑝 = (𝑋𝑝 + 𝑎𝑋𝑝 √ 𝑃 𝑠 )2,
𝑝
𝑄𝑝 = (𝑋𝑝 + 𝑎√𝛾𝑃𝑠 )2 ,
𝑄𝑝 = (√𝑃𝑃 + 𝑎√𝛾𝑃𝑠 )2,
(4.9)
(4.10) [𝑋𝑝 = √𝑃𝑝 ]
(4.11)
19
where 𝑄𝑝 is the total signal power at the PU receiver.
At PU receiver, total noise at the receiver will be the addition of normalized
gaussian noise 1, and noise due to the secondary transmission. So total noise at PU
receiver is given by:
Np = (1 + (𝑎𝑋̌𝑠 )2
(4.12)
̃𝑠 is the
Where Np is the total noise at the PU receiver. a is path loss, 𝑋
codeword that carries SU packet.
Achievable rate for Primary User can be calculated using the formula
𝑅𝑝 = 𝑙𝑜𝑔(1 + 𝑆𝐼𝑁𝑅)
(4.13)
Using Eq (4.11) and Eq (4.12) we can have above equation as
𝑅𝑝 = 𝑙𝑜𝑔(1 +
(√𝑃𝑃 +𝑎√𝛾𝑃𝑠 )2
)
(1 + (𝑎𝑋̌𝑠 )2
(4.14)
This equation can be finally used to determine the achievable rate of the
primary user while co-access with secondary user. However the value of 𝛾 should be
chosen in such a way that SINR of the primary user increases as well as rate of
secondary user is satisfactory to them.
4.1.4 At SU receiver
The received signal at SU receiver will be the sum of signal transmitted by SU
and the code word transmitted by PU. Received code word will be:
̃𝑠 + 𝑋𝑝 √𝛾𝑃𝑠 + 𝑏𝑋𝑝
𝑆𝑠 = 𝑋
𝑃
𝑝
(4.15)
20
̃𝑠 and SU receiver non causaully knows that
The desired codeword is 𝑋
𝛾𝑃
interference to the SU receiver would be 𝑋𝑝 √ 𝑃 𝑠 + 𝑏𝑋𝑝 . This is cancelled by DPC i.e
𝑝
𝛾𝑃
coding is done in such a way that √ 𝑃 𝑠 will be cancelled by 𝑏𝑋𝑝 . This is already
𝑝
stated in the paper published in [10] which shows that DPC will be success to cancel
the interference. So only the normalized noise remains in SU receiver.
Achievable rate can be given as
𝑅𝑠 = 𝑙𝑜𝑔(1 + 𝑆𝐼𝑁𝑅) ,
(4.16)
̃𝑠 |2 ) ,
𝑅𝑠 = 𝑙𝑜𝑔(1 + |𝑋
(4.17)
Using Eq (4.6) in Eq (4.17) we can derive
𝑅𝑠 = 𝑙𝑜𝑔(1 + (1 − 𝛾) 𝑃𝑠 ).
(4.18)
This equation can be used to determine the achievable rate of secondary user
when SU co-access with the PU.
4.2 Co-Access Incentive
Without SU the SINR of the Primary user will be given as
SINR = 𝑃𝑝 /1,
(4.19)
with SU, changed SINR is given as
SINR =
(√𝑃𝑃 +𝑎√𝛾𝑃𝑠 )2
(1 + (𝑎𝑋̌𝑠 )2
,
(4.20)
For primary user to be incentivized,
21
2
(√𝑃𝑃 +𝑎 √𝛾𝑃𝑠 )
(1 + (𝑎𝑋̌𝑠 )2
≥ 𝑃𝑝 + 𝐾,
(4.21)
where K is the PU co-access incentive in terms of increased SINR of the
Primary user. After some manipulation in above Equation we can derive
2
√((𝑃𝑃 +𝐾)(1−𝑃𝑃 +𝑎2 𝑃𝑠 (𝑃𝑃 +𝐾+1)))− √𝑃𝑠
𝛾≥(
𝑎√𝑃𝑠 ( 𝑃𝑝 +𝐾+1)
) .
(4.22)
This Equation can be used to calculate the portion of power to be used to relay
for given amount of PU co-access incentive to the PU user.
4.3 Acceptable SU SINR
Let λ be the minimum SINR that is desired to be received in SU receiver.
Then we can write
((1 − 𝛾) 𝑃𝑠 ) ≥ λ
(4.23)
4.4 Region Of Co-Access
If the PU co-access incentive K is not able to be offered by the SU, then the
PU does not allow the SU to co-access the licensed spectrum with it. Therefore it is
necessary to be able to find an area within the PU network that if the SU is located
within it, it would be able to provide enough incentive for co-access. In contrast to
that while calculating region of co-access acceptable SU SINR should also be
guaranteed for given 𝛾 which is used to incentivize PU by amount K. The figure 4.2
shows the basic architecture to determine the region of co-access. Here three PU’s are
assumed and PU2 relays the message of PU1 to the PU3. A SU is also intended to
22
locate around PU2 such that it increase received SINR of PU3. For this, SU should
able to receive the code word from the PU1, Thus
𝑠2 ≤ 𝑟2
(4.24)
Thus Eq (4.22) explain and finds the value of power split to provide enough
incentive to the primary user. So it guarantees that PU3 is incentivized. Eq (4.23) can
be used to assure that the rate of SU for above incentivizing is satisfactory to SU or
not. Thus combining these constraints in the below mentioned architecture we can
find the bound for the region of co-access.
PU1
r
PU2
s
a
PU3
b
SU
Figure 4.2: Architecture to determine region of co-access consisting three PU node
and a SU node
23
CHAPTER FIVE
SIMULATION RESULTS AND
INTERPRETATION
24
5.1 Simulation Parameters
Simulation is done in Matlab version r2013a. It is assumed prior to the
simulation that the spectrum sensing portion, obtaining side information has already
been done. So this simulation focuses on the achievable rates, incentives rather than
sensing of spectrum, since lots of research have been performed in case of spectrum
sensing techniques and its optimization. The power level for GSM 900 that is used in
this thesis is shown below.
Power Level
Number
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Power Output level dBm for 900
MHz
39
37
35
33
31
29
27
25
23
21
19
17
15
13
11
9
7
5
Power level in
Watt
7.943282347
5.011872336
3.16227766
1.995262315
1.258925412
0.794328235
0.501187234
0.316227766
0.199526231
0.125892541
0.079432823
0.050118723
0.031622777
0.019952623
0.012589254
0.007943282
0.005011872
0.003162278
Table 5.1: Table showing values of transmit power level of base station in dBm and
watt for GSM 900[14].
25
5.2 Achievable Rates
It is well known that achievable rate varies with the portion of SU power that is used
to relay the PU packet. So using Eq (4.14) and Eq (4.18) following results is drawn.
Figure 5.1 Plot of maximum achievable PU rate with increasing 𝛾 while PU
transmitting 5 watt and SU transmitting 7 watt.
Figure 5.1 show that the variation in Achievable PU rate with increasing 𝛾 (power
split). when 𝛾 =0 (SU does not assist PU) , then the rate is very low, it is because the
transmission of SU totally act as an interference to the PU so this reduces the
achievable rate of PU to large extent. As the value of 𝛾 goes on increasing the
achievable rate goes on increasing, at certain value of 𝛾 , the rate will be more than it
was without co-accessing. This scenario is called incentivized.
26
Figure 5. 2 Plot of maximum achievable SU rate with increasing 𝛾
while PU transmitting 5 watt and SU transmitting 7 watt.
Figure 5.2 shows the Achievable SU rate with change in 𝛾. Here when whole power
of SU is used to transmit its own code word, then achievable rate of SU is very high,
but as we decrease the portion of power then rate of SU decreases with increase in 𝛾.
27
Figure 5.3 Figure showing the comparative change in maximum
achievable rate of PU and SU with change in 𝛾
In Figure 5.3 we can see the achievable rate of both PU and SU with increasing 𝛾.
This graph is very much useful to determine that both SU and PU are satisfied with
given power split or not. Here we can see at 𝛾 = 0, the rate of PU is not zero whereas
when 𝛾 = 100 the rate of SU is 0, this is because even if SU does not assist PU, there
exist small rate for PU due to its own transmission but when all power is given to PU,
SU totally act as a repeater for PU so rate of SU goes down to zero. As the value of 𝛾
goes on increasing the rate of PU goes on increasing and rate of SU goes on
decreasing. For co-access the value of 𝛾 should be chosen in such a way that there
should be win win situation for both primary as well as secondary user. This power
split should be chosen in proper manner for the co-access.
28
Figure 5.4 shows the various curves; in which each curve represent the variation in
achievable SU rate with fixed PU power and varying SU power. PU transmit power is
assumed as 5 watt and SU transmit power ranges from 1 to 7 watt. Each curve
associated with it is plotted in the graph. The maximum achievable rate is increased
with increase in SU transmit power. This is due to more power for SU codeword as
SU tranmit power is increased.
Figure 5.4 Effect of change in transmit power level of SU on maximum achievable
rate of SU network
Figure 5.5 shows the various curves; in which each curve represent the variation in
achievable PU rate with fixed PU power and varying SU power. PU transmit power is
assumed as 5 watt and SU transmitting with power ranging from 1 to 7 watt. Each
curve associated with it is plotted in the graph. The maximum achievable rate of SU
is increased with increase in SU transmit power.
29
Figure 5.5 Effect of change in transmit power level of SU on maximum
achievable rate of PU network
5.3 Region of Co-Access
Region of co-access is defined as the geographical location around the primary user
where secondary can be located and incentivize the Primary user also finding some
space for itself. The basic architecture and methodology of how to calculate the
region of co-access is already discussed in chapter 4. Figure 4.1 shows the basic
architecture for determination of region of co-access. Thus using the equations and
the architecture with various simulation parameters the below result is obtained and
interpreted.
30
Figure 5.6: Region of co-access where SU can located between any two PU node
Figure 5.6 shows the region of co-access around the second primary user,
where SU can be located to co-access with the primary user. It consist of two circles,
the first one denotes the coverage of the PU1, and the second one denotes the
coverage of PU3, PU2 is repeater, SU should be within the first circle so that it can
relay the message of PU for incentivizing, but this location should be proper enough
that PU3 gets the relayed message with given incentive. Thus the region between the
two circles is what we called the region of co-access.
31
CHAPTER SIX
EPILOGUE
32
6.1 Limitations
The various limitations of this dissertation are as follows:
1. For now, fixed transmit power is considered for simulation purpose.
2. The spectrum is assumed that it is sensed and priori information of PU is
already given to the SU.
6.2 Future Enhancements
This possibility of future enhancement are as follows:
1. It can be used for all types of wireless bands.
2. Adaptive power control according to PU power can be implemented in SU
transmitted.
3. System modeling with numbers of node pair can be done.
6.3 Schedule
This thesis is aimed to complete within twenty two weeks.
Collection and Study of Materials
Familiarization of N/Wsimulator…
Algorithm Study, Design and…
System Design
Integration and Testing
Output Analysis
Documentation and Report
0
5
10
15
20
25
Table 6.1 Time schedule
33
6.4 Conclusion
This dissertation presents a dynamic spectrum access architecture termed
DSCA. DSCA enables SUs to co-access spectrum with PUs, i.e., simultaneously
transmit with PUs. This significantly reduces the disruption to SU communication
due to the Resurgence of PU traffic. Furthermore, it offers guaranteed incentives to
PUs to allow the co-access of SUs, as well as guaranteed performance for SUs in
spectrum co-access. Together, both PUs and SUs benefit from the DSCA architecture.
We have defined the co-access incentives for both PUs and SUs, and derived a model
to compute the region of co-access based on the co-access incentives.
34
REFERENCES
35
[1] S.Sun.,Y. Ju.,“overlay cognitive radio ofdm system for 4g cellular network” ,
IEEE Wireless Communication, April 2013.
[2] A. Goldsmith et al., “Breaking spectrum gridlock with cognitive radios: an
information theoretic perspective,” Proc. IEEE, vol. 97, no. 5, May 2009, pp.
894–914.
[3] C. Xin, M. Song, L. Ma, G. Hsieh, and C.-C. Shen, “An incentivized
cooperative architecture for dynamic spectrum access networks,” IEEE Trans.
Wireless Commun., vol. 12, no. 10, pp. 5154–5161, Oct. 2013.
[4] J. Mitola, “Cognitive radio,” Licentiate proposal, KTH, Stockholm,
Sweden,Dec. 1998.
[5] H. Bezabih et al., “Digital broadcasting: increasing the available white space
spectrum using tv receiver information,” IEEE Vehic. Tech. Mag., vol. 7 no.
1, Mar. 2012, pp. 24–30.
[6] K. Shin, H. Kim, A. Min, and A. Kumar, “cognitive radios for dynamic
spectrum access: from concept to reality,” wireless communications, IEEE,
vol. 17, no. 6, pp. 64–74, 2010.
[7] M. Song, C. Xin, Y. Zhao, and X. Cheng, “Dynamic spectrum access: from
cognitive radio to network radio” , Wireless Communications, IEEE,vol. 19,
no. 1, pp. 23–29, 2012.
[8] M. Costa, “Writing in a dirty paper (corresp.)”, IEEE transaction on
Information Theory, Vol 29, No.3, pp. 439-441, 1983.
36
[9] X. Guan, Y. Cai, Y. Sheng, and W. Yang, “Exploiting primary retransmission
to improve secondary throughput by cognitive relaying with best-relay
selection,” IET Communications, vol. 6, no. 12, pp. 1769–1780, 2012.
[10] M. Mohammadin, S. seltine., “Joint interference cancellation and dirty paper
coding for cog- nitive cellular networks,” in Wireless Communications and
Networking Conference (WCNC), 2011 IEEE. IEEE, 2011, pp. 1972–1976.
[11]
R.
Zhang,
Member,
IEEE,
and
Y-C
Liang,
Senior
Member,
IEE.,“Investigation on Multiuser Diversity in Spectrum Sharing Based
Cognitive Radio Networks”.
[12] Federal Communications Commission Spectrum Policy Task Force, “Report
of the Spectrum Efficiency Working Group,” Technica Report 02-135, no.
November,
2002.
Download
available
at
http://www.fcc.gov/sptf/files/SEWGFinalReport 1.pdf
[13] T. cover, “Broadcast Channels”, IEEE transaction on Information Theory,
Vol 18, pp. 2-14, 1972.
[14] Transmit power level of GSM base station. Download
available at
http://www.radioelectronics.com/info/cellulartelecomms/gsm_technical/po
wer.php
37