Location Information Based Interference Control for Cognitive Radio ...
2013 IEEE Wireless Communications and Networking Conference (WCNC): PHY
Location Information Based Interference Control for Cognitive Radio Network in TV White Spaces Xin Tao∗† , Zhifeng Zhao∗† , and Honggang Zhang∗†
∗ York-Zhejiang
Lab for Cognitive Radio and Green Communications of Information Science and Electronic Engineering Zhejiang University, Zheda Road 38, Hangzhou 310027, China Email: {taoxin, zhaozf, honggangzhang}@zju.edu.cn † Dept.
Abstract—Controlling the interference from secondary users utilizing TV white spaces (TVWS) to protect primary users (PUs) is of vital importance in multicarrier based cognitive radio (CR) systems. Due to the complexity in the real implementation scenarios, regulations relative to TV white spaces may not be fully implemented, and even meeting all the regulatory requirements can’t guarantee that the primary users are not influenced completely. In this paper, we propose a power allocation algorithm based on location information of the secondary users to minimize the aggregate interference to the licensed TV receivers at the border of TV service contour while making the capacity of secondary system reach a certain level. In the meantime, we need to assure that the minimum aggregate interference meets the required interference protection ratios of incumbent users. Since the use of geolocation and database access is mandated by the regulatory authorities, it is quite easy for the white space database (WSDB) to obtain location information of the secondary users and allocate their power accordingly. Numerical simulations are carried out to validate the effectiveness of the proposed interference control algorithm.
I. I NTRODUCTION Allowing unlicensed cognitive radio (CR) access to the geographically unused spectrum in TV broadcasting band, known as TV white spaces (TVWS), will greatly increase the efficiency of the spectrum usage. This kind of dynamic spectrum operation for cognitive radio in TVWS has been authorized by the Federal Communications Commission (FCC) in USA since 2008 [1]. Similar to USA, the digital TV switchover program is expected to be completed in 2012 in UK, which will make more TV channels unused in different geographical locations. According to the survey [2], for most of the UK locations, there will be around 100MHz interleaved spectrum available while the averaged capacity per location can be over 150MHz. Moreover, the relatively low UHF band of TVWS has a good propagation characteristic, which has been verified to outperform the 5GHz and 2.4GHz bands [3].That’s why unlicensed access to the TVWS has caused so much attention in recent years around the world. However, the premise of utilizing TVWS is that unlicensed spectrum access on a secondary basis should not produce any harmful interference to the incumbent users. Since the geolocation database techniques have a good effect in protecting the incumbent users [1] [4], regulatory authorities have mandated the use of this technology. After obtaining the location information of white space devices (WSD), white space database
978-1-4673-5939-9/13/$31.00 ©2013 IEEE
(WSDB) can inform WSD of available channels or permitted maximum transmit power level to control interference induced by WSD to the incumbent users under a certain level. In [5], the secondary cognitive radio system’s aggregate interference to a TV receiver at the TV service contour’s border has been studied. According to the regulation of interference protection ratios specified in FCC’s report [1], the authors in [5] derived an interference threshold that a TV receiver at the TV service contour’s border could tolerate at most. Then the solution of the corresponding optimization problem became the well-known water-filling algorithm. But the scheme in [5] was too brief to give a final specific result. In [6], an interference control scheme, which is aiming at keeping multiple secondary systems’ aggregate interference at the victim TV receivers under specific interference limits, has been proposed. And M.Shaat et al. in [7] proposed a relatively complicated power allocation algorithm called PI-Algorithm in cognitive radio networks. The point is, however, that studies mentioned above just focus on the optimization problem under interference constraints. Considering the implementation complexity in the reality, situations may change from one place to another and the regulations relative to the TV white spaces may not be fully implemented. Because of the difference, some requirements are not suitable for all the cases and sometimes even meeting the interference threshold can’t guarantee that the primary users (PUs) are not affected completely. In this paper, an interference control algorithm based on location information of the secondary users to minimize the aggregate interference to the TV receivers at the TV service contour’s border while making the capacity of secondary system reach a certain level is proposed. By selecting different secondary systems’ capacity level, we can minimize the aggregate interferences and control them under different interference limits to be adaptive to different situations. Moreover, our proposed scheme makes sure that the minimum aggregate interference meets the required interference protection ratio of incumbent users. The rest of this paper is organized as following: In Section II, the channel and system model is described and Section III formulates the problem and presents the power allocation algorithm. Numerical results are shown in Section IV and finally, the paper is concluded in Section V.
3614
be expressed as
II. C HANNEL AND S YSTEM M ODEL We consider the uplink scenario in TVWS systems. As shown in Fig. 1, the CR system is located adjacent to the TV service contour with a separation distance in between. The separation distance, which is to protect the TV receivers from the potential harmful interference, has a different value when the unlicensed users, also referred to as secondary users (SUs), operate on the co-channel or adjacent channel bands of TV white spaces. What’s more, the separation distance between a particular WSD and a TV station’s service contour depends mainly on the height of the device’s antenna above ground, according to the FCC’s report [1]. Considering that the TV white space system’s coverage area is relatively larger as defined in the IEEE 802.22 standard [8], large-scale path loss is dominant and we use path loss channel model as follows 2
gSU (ri ) = const · (ht hr ) ri−αSU 2
gT V (R) = const · (hT hR ) R
−αT V
di +B/2
Ii =
(3)
di −B/2
where gi is the channel gain between the ith secondary user and the PU receiver. di is the spectral distance between the ith secondary user band and the PU band. Φi is the power spectrum density (PSD) of the ith secondary user signal. Besides, the expression of the PSD depends on the adopted multicarrier technique, OFDM or FBMC [10]. Ωi denotes the interference factor of the ith secondary user. We can see that once a multicarrier technique is adopted, Ωi is mainly associated with gi . Similarly, the interference induced by the PU signal into the signal band of the ith secondary user is di +B SU /2
(1)
2 |yi | Ψ ejw dw
Ji =
(2)
where ht , hr are the heights of the transmit antenna and the receive antenna of the secondary users in white space system, respectively and similarly, hT , hR are the heights of the TV transmitter antenna and TV receiver antenna. R denotes the radius of the TV service area, ri denotes the distance between the ith secondary user and the CR base station while αT V , αSU are the propagation path loss exponents for TV and secondary transmission, respectively.
2
|gi | Φi (f )df = Pi Ωi
(4)
di −BSU /2
where Ψ ejw is the PSD of the PU transmit signal. Without losing generality, we consider the representative case of having a single TV receiver at the TV protected contour’s border. Thus, the aggregate interference from the N secondary users ISU at the TV receiver is ISU =
N
Pi Ω i
(5)
i=1
For the TV receiver, the signal-to-interference-and-noise ratio (SINR), γT V , can be expressed as γT V =
79VHUYLFHFRQWRXU
79 UHFHLYHU
(6)
where X = PT V · gT V denotes the desired TV signal power, PT V denotes the TV’s transmission power, gT V is the channel model of the TV transmission system as mentioned earlier, ISU is the aggregate interference from the secondary users, and PN is the noise power. To ensure quality of service (QoS), the TV receiver has to satisfy at least a target SINR value, γ, which is called interference protection ratio in the FCC’s report [1] specified in terms of desire-to-undesired (D/U) signal levels
79WUDQVPLWWHU
6HFRQGDU\ 8VHU 68
&5EDVHVWDWLRQ
6HSDUDWLRQGLVWDQFH
Fig. 1.
X ISU + PN
γT V ≥ γ
System model.
In the multicarrier based cognitive radio system, the cognitive base station (CBS) transmits to its secondary users and generates potential interference to the victim TV receivers. In the meantime, the TV transmitter may interfere with the secondary user as well. We presume that the frequency spectrum of the CR system is divided into N groups of subcarriers with each secondary user having one group of subcarriers. The interference introduced by the ith secondary user to the PU (i.e. TV receiver) band, Ii , is the integration of the ith secondary user signal’s PSD across the PU band, B, and can
(7)
The interference protection ratio of digital TV is 23dB for cochannel, -28dB for lower adjacent channel and -26dB for upper adjacent channel, according to the FCC’s report [1]. Thus, the aggregate interference threshold, Ith , can be determined as ISU ≤
X − PN = Ith γ
(8)
Later in Section IV, it will be shown that the minimum aggregate interference due to secondary transmission needs to be controlled under this aggregate interference threshold that is calculated by the above equation.
3615
III. P ROBLEM F ORMULATION AND P OWER C ONTROL In this section, we will formulate the constraint optimization problem for the CR system and present our proposed power control algorithm. For integrity of the derivation and analysis, the formulation of the problem will be done for the downlink firstly and then be simplified for the uplink. Recall that, our objective is to minimize the aggregate interference to a TV receiver at the border of TV service contour subject to the total capacity requirement and total transmit power constraint of the CR system. Therefore, the optimization problem can be formulated as follows P 1 : min Pi
N
Pi Ωi
i=1
Subject to N
(9)
Ci ≥ CT
i=1 N
Pi ≤ PT ; Pi ≥ 0
where Pi is the transmit power of the ith secondary user, and Ωi is the interference factor of the ith secondary user. Moreover, Ci is the capacity of the ith secondary user and CT is the total capacity threshold of the CR system, PT is the total power budget while N is the number of secondary users. For the sake of brevity, we presume that all secondary users have the same bandwidth.Using the Shannon capacity 2 formula Ci = log2 1 + Pi |hi | σ 2 , we can get Pi =
|hi |
2
2
Ci
−1
Ci
(10)
N σ 2 Ci − 1 Ωi 2 2 i=1 |hi |
Ci ≥ CT
i
i=1
where [x] = max (0, x). Proof. Rearranging the last condition in (13), we can obtain 2 | α|h i Ci∗ = log 2 (15) σ 2 (Ωi + β − μi ) Substituting (15) into (10), we have Pi∗ =
α σ2 − 2 Ωi + β − μi |hi |
Considering μi Pi∗ = 0, if Pi∗
=
2
σ2 |hi |2
Location Information Based Interference Control for Cognitive Radio Network in TV White Spaces Xin Tao∗† , Zhifeng Zhao∗† , and Honggang Zhang∗†
∗ York-Zhejiang
Lab for Cognitive Radio and Green Communications of Information Science and Electronic Engineering Zhejiang University, Zheda Road 38, Hangzhou 310027, China Email: {taoxin, zhaozf, honggangzhang}@zju.edu.cn † Dept.
Abstract—Controlling the interference from secondary users utilizing TV white spaces (TVWS) to protect primary users (PUs) is of vital importance in multicarrier based cognitive radio (CR) systems. Due to the complexity in the real implementation scenarios, regulations relative to TV white spaces may not be fully implemented, and even meeting all the regulatory requirements can’t guarantee that the primary users are not influenced completely. In this paper, we propose a power allocation algorithm based on location information of the secondary users to minimize the aggregate interference to the licensed TV receivers at the border of TV service contour while making the capacity of secondary system reach a certain level. In the meantime, we need to assure that the minimum aggregate interference meets the required interference protection ratios of incumbent users. Since the use of geolocation and database access is mandated by the regulatory authorities, it is quite easy for the white space database (WSDB) to obtain location information of the secondary users and allocate their power accordingly. Numerical simulations are carried out to validate the effectiveness of the proposed interference control algorithm.
I. I NTRODUCTION Allowing unlicensed cognitive radio (CR) access to the geographically unused spectrum in TV broadcasting band, known as TV white spaces (TVWS), will greatly increase the efficiency of the spectrum usage. This kind of dynamic spectrum operation for cognitive radio in TVWS has been authorized by the Federal Communications Commission (FCC) in USA since 2008 [1]. Similar to USA, the digital TV switchover program is expected to be completed in 2012 in UK, which will make more TV channels unused in different geographical locations. According to the survey [2], for most of the UK locations, there will be around 100MHz interleaved spectrum available while the averaged capacity per location can be over 150MHz. Moreover, the relatively low UHF band of TVWS has a good propagation characteristic, which has been verified to outperform the 5GHz and 2.4GHz bands [3].That’s why unlicensed access to the TVWS has caused so much attention in recent years around the world. However, the premise of utilizing TVWS is that unlicensed spectrum access on a secondary basis should not produce any harmful interference to the incumbent users. Since the geolocation database techniques have a good effect in protecting the incumbent users [1] [4], regulatory authorities have mandated the use of this technology. After obtaining the location information of white space devices (WSD), white space database
978-1-4673-5939-9/13/$31.00 ©2013 IEEE
(WSDB) can inform WSD of available channels or permitted maximum transmit power level to control interference induced by WSD to the incumbent users under a certain level. In [5], the secondary cognitive radio system’s aggregate interference to a TV receiver at the TV service contour’s border has been studied. According to the regulation of interference protection ratios specified in FCC’s report [1], the authors in [5] derived an interference threshold that a TV receiver at the TV service contour’s border could tolerate at most. Then the solution of the corresponding optimization problem became the well-known water-filling algorithm. But the scheme in [5] was too brief to give a final specific result. In [6], an interference control scheme, which is aiming at keeping multiple secondary systems’ aggregate interference at the victim TV receivers under specific interference limits, has been proposed. And M.Shaat et al. in [7] proposed a relatively complicated power allocation algorithm called PI-Algorithm in cognitive radio networks. The point is, however, that studies mentioned above just focus on the optimization problem under interference constraints. Considering the implementation complexity in the reality, situations may change from one place to another and the regulations relative to the TV white spaces may not be fully implemented. Because of the difference, some requirements are not suitable for all the cases and sometimes even meeting the interference threshold can’t guarantee that the primary users (PUs) are not affected completely. In this paper, an interference control algorithm based on location information of the secondary users to minimize the aggregate interference to the TV receivers at the TV service contour’s border while making the capacity of secondary system reach a certain level is proposed. By selecting different secondary systems’ capacity level, we can minimize the aggregate interferences and control them under different interference limits to be adaptive to different situations. Moreover, our proposed scheme makes sure that the minimum aggregate interference meets the required interference protection ratio of incumbent users. The rest of this paper is organized as following: In Section II, the channel and system model is described and Section III formulates the problem and presents the power allocation algorithm. Numerical results are shown in Section IV and finally, the paper is concluded in Section V.
3614
be expressed as
II. C HANNEL AND S YSTEM M ODEL We consider the uplink scenario in TVWS systems. As shown in Fig. 1, the CR system is located adjacent to the TV service contour with a separation distance in between. The separation distance, which is to protect the TV receivers from the potential harmful interference, has a different value when the unlicensed users, also referred to as secondary users (SUs), operate on the co-channel or adjacent channel bands of TV white spaces. What’s more, the separation distance between a particular WSD and a TV station’s service contour depends mainly on the height of the device’s antenna above ground, according to the FCC’s report [1]. Considering that the TV white space system’s coverage area is relatively larger as defined in the IEEE 802.22 standard [8], large-scale path loss is dominant and we use path loss channel model as follows 2
gSU (ri ) = const · (ht hr ) ri−αSU 2
gT V (R) = const · (hT hR ) R
−αT V
di +B/2
Ii =
(3)
di −B/2
where gi is the channel gain between the ith secondary user and the PU receiver. di is the spectral distance between the ith secondary user band and the PU band. Φi is the power spectrum density (PSD) of the ith secondary user signal. Besides, the expression of the PSD depends on the adopted multicarrier technique, OFDM or FBMC [10]. Ωi denotes the interference factor of the ith secondary user. We can see that once a multicarrier technique is adopted, Ωi is mainly associated with gi . Similarly, the interference induced by the PU signal into the signal band of the ith secondary user is di +B SU /2
(1)
2 |yi | Ψ ejw dw
Ji =
(2)
where ht , hr are the heights of the transmit antenna and the receive antenna of the secondary users in white space system, respectively and similarly, hT , hR are the heights of the TV transmitter antenna and TV receiver antenna. R denotes the radius of the TV service area, ri denotes the distance between the ith secondary user and the CR base station while αT V , αSU are the propagation path loss exponents for TV and secondary transmission, respectively.
2
|gi | Φi (f )df = Pi Ωi
(4)
di −BSU /2
where Ψ ejw is the PSD of the PU transmit signal. Without losing generality, we consider the representative case of having a single TV receiver at the TV protected contour’s border. Thus, the aggregate interference from the N secondary users ISU at the TV receiver is ISU =
N
Pi Ω i
(5)
i=1
For the TV receiver, the signal-to-interference-and-noise ratio (SINR), γT V , can be expressed as γT V =
79VHUYLFHFRQWRXU
79 UHFHLYHU
(6)
where X = PT V · gT V denotes the desired TV signal power, PT V denotes the TV’s transmission power, gT V is the channel model of the TV transmission system as mentioned earlier, ISU is the aggregate interference from the secondary users, and PN is the noise power. To ensure quality of service (QoS), the TV receiver has to satisfy at least a target SINR value, γ, which is called interference protection ratio in the FCC’s report [1] specified in terms of desire-to-undesired (D/U) signal levels
79WUDQVPLWWHU
6HFRQGDU\ 8VHU 68
&5EDVHVWDWLRQ
6HSDUDWLRQGLVWDQFH
Fig. 1.
X ISU + PN
γT V ≥ γ
System model.
In the multicarrier based cognitive radio system, the cognitive base station (CBS) transmits to its secondary users and generates potential interference to the victim TV receivers. In the meantime, the TV transmitter may interfere with the secondary user as well. We presume that the frequency spectrum of the CR system is divided into N groups of subcarriers with each secondary user having one group of subcarriers. The interference introduced by the ith secondary user to the PU (i.e. TV receiver) band, Ii , is the integration of the ith secondary user signal’s PSD across the PU band, B, and can
(7)
The interference protection ratio of digital TV is 23dB for cochannel, -28dB for lower adjacent channel and -26dB for upper adjacent channel, according to the FCC’s report [1]. Thus, the aggregate interference threshold, Ith , can be determined as ISU ≤
X − PN = Ith γ
(8)
Later in Section IV, it will be shown that the minimum aggregate interference due to secondary transmission needs to be controlled under this aggregate interference threshold that is calculated by the above equation.
3615
III. P ROBLEM F ORMULATION AND P OWER C ONTROL In this section, we will formulate the constraint optimization problem for the CR system and present our proposed power control algorithm. For integrity of the derivation and analysis, the formulation of the problem will be done for the downlink firstly and then be simplified for the uplink. Recall that, our objective is to minimize the aggregate interference to a TV receiver at the border of TV service contour subject to the total capacity requirement and total transmit power constraint of the CR system. Therefore, the optimization problem can be formulated as follows P 1 : min Pi
N
Pi Ωi
i=1
Subject to N
(9)
Ci ≥ CT
i=1 N
Pi ≤ PT ; Pi ≥ 0
where Pi is the transmit power of the ith secondary user, and Ωi is the interference factor of the ith secondary user. Moreover, Ci is the capacity of the ith secondary user and CT is the total capacity threshold of the CR system, PT is the total power budget while N is the number of secondary users. For the sake of brevity, we presume that all secondary users have the same bandwidth.Using the Shannon capacity 2 formula Ci = log2 1 + Pi |hi | σ 2 , we can get Pi =
|hi |
2
2
Ci
−1
Ci
(10)
N σ 2 Ci − 1 Ωi 2 2 i=1 |hi |
Ci ≥ CT
i
i=1
where [x] = max (0, x). Proof. Rearranging the last condition in (13), we can obtain 2 | α|h i Ci∗ = log 2 (15) σ 2 (Ωi + β − μi ) Substituting (15) into (10), we have Pi∗ =
α σ2 − 2 Ωi + β − μi |hi |
Considering μi Pi∗ = 0, if Pi∗
=
2
σ2 |hi |2
