Relay Selection and Resource Allocation in LTE-Advanced Cognitive Relay Networks

Transcription

1 International Journal on Communication Antenna and Propagation (I.Re.C.A.P.), Vol. 1, N. 4 Augut 2011 Relay Selection and Reource Allocation in LTE-Advanced Cognitive Relay Network Ardalan Alizadeh, Seyed Mohammad-Sajad Sadough, Seyed Ali Ghorahi Abtract In thi work, we conider the problem of joint relay election and reource allocation in Long Term Evolution-Advanced (LTE-A) ytem which provide both cognitive radio and relaying cheme capabilitie. We aume that the total network i deployed in an overlay cheme where the primary uer communicate via a relay aited LTE-A network, ome of the econdary uer play the relaying role and the remained node are communicate by centralized network model in the licened pectrum. In the firt tep of the propoed procedure, the cognitive radio bae tation (CBS) elect the higher gain component carrier (CC) channel and allocate one CC to each econdary terminal which i denoted a cutomer premie equipment (CPE). The power updating algorithm i provided in the econd tep of the propoed cheme which give the maximum SINR in the econdary CPE while keeping the minimum SINR threhold at the primary receiver. In the third tep of the propoed algorithm, CC are re-allocated to cognitive radio CPE. Simulation reult are provided to compare the performance of our propoed relay election and reource allocation algorithm with random relay election and uniform power allocation repectively. Copyright 2011 Praie Worthy Prize S.r.l. - All right reerved. Keyword: Cognitive Radio Network, Dynamic Power Allocation, LTE-Advanced, Relay Selection Nomenclature CR Cognitive radio LTE-A Long Term Evolution-Advanced CC Component carrier CPE Cutomer premie equipment CBS Cognitive radio bae tation UE Uer equipment Evolved Node B c Component carrier channel M Number of CPE N Number of CC K Number of CR-CPE f c,i Channel gain between the and the i-th CPE on channel c f Channel gain between, and CBS c,0 g c,i Channel gain between the UE and the i-th CPE on component carrier c g Channel gain between UE and CBS c,0 h c,i,j Channel gain between the i-th and the j-th CPE on channel c P Tranmit power of on channel c c P Tranmit power of UE on channel c UE c c, 0 P Tranmit power of the CBS on channel c I. Introduction Cognitive radio (CR) i a new deign paradigm to combat the problem of carce and expenive pectrum reource in wirele communication [1]. The baic idea of CR i to allow a econdary (unlicened) uer to utilize a frequency band already allocated to primary (licened) uer. Recently, variou approache are uggeted to mitigate the effect of econdary uer on licened network [2], [3]. In the opportunitic or o-called interweave approach [2], the econdary uer ha to ene the pectrum contantly in order to detect the pectral hole or white pace before tranmitting it own ignal. Obviouly, the cognitive radio uer hould give back the pectrum once the preence of the primary uer i detected in order to minimize their harmful interference to licened uer. Rather than detecting white pace, in underlay and overlay approache, cognitive econdary uer and primary uer() tranmit imultaneouly, while econdary uer ue their cognitive capabilitie to control the amount of interference upon the primary uer(). The underlay approach i imilar to ultra wideband (UWB) ytem which enforce the pectral mak for econdary uer to hold the impoed interference below a predefined threhold. In the overlay approach, the econdary uer hare part of it power reource with the primary uer to provide a relay-aited tranmiion. Manucript received and revied July 2011, accepted Augut 2011 Copyright 2011 Praie Worthy Prize S.r.l. - All right reerved 303

2 A. Alizadeh, S. M.-S. Sadough, S. A. Ghorahi Therefore, the econdary network compenate the impoed interference by increaing the ignal-tointerference-plu-noie ratio (SINR) of primary receiver. Then, the baic idea of overlay approach i to allocate power and channel reource to whole network, while utilizing the requirement of primary uer() concurrently. In thi technique, channel tate information (CSI) hould be known in two network. In uch overlay relaying cheme, the power efficiency i a critical iue for common relay node which erve ingle or multiple uer pair to communicate. Some relay power allocation trategie a well a other reource allocation algorithm are currently propoed in the literature [4]-[7] for relay-aited communication. Alo, uing the cognitive radio capabilitie in uni-directional relaying network ha been tudied for different application [8]-[10]. In [11], [12], different relay election algorithm are propoed which are baed on SNR maximization of the primary network. In [13], the model i introduced in which the cognitive radio network hare it mobile device with the primary network to increae the probability of licened pectrum hole uage by opportunitic approach. In thi paper, the model i baed on the overlay approach in which elected econdary uer ait (relay) the primary network in order to utilize the licened pectrum and the remained econdary uer are communicating with a cognitive bae tation (CBS). In uch overlay cheme, unlike the opportunitic one, cognitive radio uer (denoted a cutomer premie equipment (CPE)) can communicate with their bae tation while they are not impoing harmful interference on the primary relay network. It i aumed that a central controller (i.e., CBS) can obtain SINR meaurement reult and know CSI among all CR node through dedicated control channel. A a cae tudy, the propoed algorithm are applied in a long term evolution- Advanced (LTE-A) baed model. LTE-A i an emerging technology provided by the third Generation Partnerhip Project (3GPP) in which new feature uch a relaying cheme [14], [15] and cognitive radio [16] have been conidered for the future thi next generation ytem. The main contribution of thi paper are a follow: 1- A novel application for cognitive relaying in LTE- Advanced ytem i introduced. Thi new cenario improve pectrum utilization and increae the SINR of the econdary network, while allowing the primary pair to communicate via a uni-directional relaying mechanim. 2- A new reource management cheme i propoed for the above mentioned cenario, in order to optimize the utilization of relaying cheme between LTE-A tranmitter and receiver Thi cheme conit of three main tage a hown in Fig. 1: relay election for each component carrier (elected CPE are denoted by RE-CPE), power allocation for primary and econdary uer, and finally channel allocation for thoe CR node that are not participating in relaying (the remained CPE are denoted by CR-CPE), in order to maximize the throughput of the econdary network. The ret of thi paper i organized a follow. In Section II, we decribe the ytem model of propoed network. In Section III, joint channel and relay election, a well a reource allocation algorithm are explained, repectively. The power updating approach and channel allocation cheme are provided in Section IV and V. Simulation reult and dicuion are preented in Section V, and finally, Section 6 conclude the paper. II. Sytem Model The 100 MHz LTE-A bandwidth conit of five component carrier (CC), each one have a bandwidth of 20 MHz [17] (Fig. 2). The feature of each CC are in coherence with LTE Releae 8. The total bandwidth of the LTE-Advanced can be conidered le than 100 MHz, and therefore may conit of up to five component carrier. The frequency band and pectrum allocation expreed via the number of component carrier and their bandwidth are configurable and known a priori by primary bae tation (hereafter denoted by according to the 3GPP terminology) and by CBS. By uing carrier aggregation, an LTE-Advanced mobile terminal (uer equipment (UE)) can be jointly cheduled on multiple component carrier providing higher data rate than conventional LTE ytem, a hown in Fig. 2. We conider a two-tep amplify-and-forward (AF) relaying for all component carrier of the LTE-Advanced network. We alo aume that there i no direct link between and UE a hown in Fig. 3. In the firt tep of the AF relaying, tranmit ingle or multiple component carrier to the M CPE, imultaneouly. In the econd tep, each CPE i able to amplify it received ignal and broadcat it to the UE a well a to the CBS. Since the bandwidth of each component carrier and the required power for relaying are uually large, we aume that each CPE can only ue one component carrier for relaying. Fig. 1. Propoed reource management procedure Copyright 2011 Praie Worthy Prize S.r.l. - All right reerved Int. Journal on Communication Antenna and Propagation, Vol. 1, N

3 A. Alizadeh, S. M.-S. Sadough, S. A. Ghorahi Fig. 2. Carrier aggregation and ingle component carrier aignment to each UE for LTE-A Rel. 8 Fig. 3. Notation of channel gain for the c-th component carrier, where 1 c 5 and M>Max(c) II.1. By thi approach, part of CPE work a relay node (according to the maximum number of CC, the number of relay will be le than five) and the remained node form a eparate econdary network for intra-econdary network communication. Fig. 3 how the block diagram of the conidered model for the c-th component carrier channel ( 1 c 5 ). In thi Figure, f c,i denote the channel gain between the Problem Formulation In the firt tep, the end information ymbol 1 to CPE. The received ignal at the i-th RE-CPE and channel c which i determined by relay election algorithm i calculated a: rc,i = Pc f c,i 1 + υi and the i-th CPE on channel c, where 0 i M, and f c,0 i the channel gain between, and CBS. (1) where υi i the noie at the i-th relay. We conider a tationary fading where the channel tay contant during reource allocation tage. A mentioned, the cognitive radio network hare N of it fixed-location CPE a relay node with the primary network (RE-CPE). Alo, the cognitive relay network ue a bae tation to control and upport CR-CPE and provide a point-to-multipoint cognitive radio network. In the econd tep, N CPE are elected to tranmit via N CC while each RE-CPE end in one CC. We aume that the j-th relay i elected and thi relay amplifie the received ignal and forward it to primary receiver. Therefore, the j-th relay end t j = α j rc, j where α j i the amplify weighting of the j-th Similarly, gc,i denote the channel gain between the UE and the i-th CPE on component carrier c, where 0 i M, and gc,0 i the channel gain between UE and CBS. Alo, hc,i, j denote the channel gain between the i-th and the j-th CPE on channel c, where 0 i, j M, and hc,i,0 i the channel gain between CBS and the i-th CPE. In thi Figure, Pc denote the tranmit power of on channel c, and PcUE denote the tranmit power of UE on channel c. Pc,0 denote the tranmit power of the CBS on c-th channel. relay. We aume contant uniform weighting for all relay. The received ignal at the econd tep in UE i: Copyright 2011 Praie Worthy Prize S.r.l. - All right reerved Int. Journal on Communication Antenna and Propagation, Vol. 1, N

4 A. Alizadeh, S. M.-S. Sadough, S. A. Ghorahi ( 1 υ ) UE c j c,j α j c,j c c,i j x = tg + n= g P f + + n (2) where n i the additive noie in the econd tep of relaying. We aume that all noie are i.i.d complex Gauian random variable with zero mean and unit variance. We alo aume that the primary and econdary network can cooperate via a common control channel. In thi model, two network can operate imultaneouly while the cognitive network doe not impoe the interference temperature above a predefined threhold. Therefore, the required SINR of primary receiver remain above a threhold level. The objective i joint power and channel allocation for econdary uer in order to maximize the total throughput of the econdary uer, while maintaining the required SINR level for the primary receiver. We aume a contant uniform tranmit power for primary network. We conider M (M 5) fixed CPE, which are randomly ditributed between primary tranmitter and receiver. In thi model, N of CPE are choen a RE-CPE by the propoed relay election algorithm where N i the number of component carrier for UE that i conidered by. Note that each of the K remained CPE (CR- CPE, K =M - N) can reue only one of N component carrier, while the SINR of primary receiver and all CPE are held above a predefined threhold. In thi paper, only the downlink cheme (from CBS to CR- CPE and to UE) ha been conidered. II.2. SINR at Receiver After the relay election approach, two ubet of node are divided from the initial et of CPE, i.e., RE- CPE and CR-CPE for the econdary network. Conider the downlink cenario in the cognitive radio network for CR CPE( 1) each channel c, let γ c,i denoting the SINR experienced by CR-CPE at firt tep of relaying a: γ CR CPE c, c,i, c,i 0 c () 1 P 0h 0 = N + P f CR CPE( 2) and γ c,i denote the SINR experienced by CR- CPE at econd tep of relaying a: γ c,i CR CPE( 2) Pc, 0hc,i, 0 c,i = N0 + α j fc,jhc,i,jpc where the j-th CPE (RE-CPE) i elected for channel c during the relay election procedure. Alo, we can write the SINR at UE a: γ Pc α j fc,jg UE c,j c,i = N0 + Pc, 0gc, 0 + Pc, 0hc,j, 0α jgc,j (3) (4) (5) Two interference are impoed to UE receiver during relaying tep. The term Pc, 0g c, 0 i the effect of CBS in the econd tep and term P c, 0α j fc,jhc,j, 0 i due to the firt tep of relaying which i amplified and forwarded to the primary receiver. For atifying a minimum performance requirement at the primary network, we aume that the received SINR at the UE mut be above a predefined value, γ p. In particular, when the cognitive radio network operate in the downlink cenario, we mut have γ UE p c γ. Finally, we can write the objective of our propoed reource management cheme a the following optimization problem: arg max K CR CPE λγ i c,i = 12, i= 1 UE p γc,j CBS Pc, 0 Pmax Pc = P ( ) ubject to : γ for j = 1,...,N In thi problem, λ i equal to 0 or 1 how the i-th CR- CPE election approach in the final tage of the reource allocation, which i applied by bipartite matching. III. Joint Relay Selection and Channel Allocation We conider a relay election cheme in the firt tep of reource allocation operation. The objective of thi tage i to elect the relay and to aign the bet channel for elected RE-CPE. A mentioned, ince the bandwidth of each component carrier i large and each relay require high tranmit power for multi-carrier tranmiion, we aume that each cognitive relay can only amplify-and-forward via one component carrier. For thi purpoe, we propoe a comparative heuritic relay election algorithm that find the bet total gain for conidered order of maximum gain. After that, the primary network can ue elected CPE a RE-CPE where channel aignment i jointly conidered. Alo, the remained node are ued a CR-CPE for the econdary network. Note that relay election algorithm i proceed in CBS. Fig. 4 illutrate the flowchart of the propoed relay election algorithm. In thi algorithm, the CBS allocate CC to the i-th CPE a follow: The iteration number, r, and CPE index, j, are initially et to 1. In the r-th iteration (r = 1,..., R), the i-th CPE, to which no component carrier ha allocated yet, compare the r-th maximum channel gain to the j-th CPE, j = 1,, M. If a given channel ha the r-th bet gain (Max r (f c,i g c,i )) for exactly one CPE (except the i-th CPE), thi channel i allocated to that CPE and the next iteration tart. (6) Copyright 2011 Praie Worthy Prize S.r.l. - All right reerved Int. Journal on Communication Antenna and Propagation, Vol. 1, N

5 A. Alizadeh, S. M.-S. Sadough, S. A. Ghorahi value of P ( ) c, 0 0 P c i kept contant., while the primary tranmitted power Power Updating: At k-th iteration, the CBS update the tranmit power a follow: P k + 1 = P k δ ( ) ( ) 0,c 0,c Here, δ i a power caling factor, which i lightly greater than one. Fig. 4. Comparative heuritic relay election algorithm to elect CPE a relay and CC aignment. If all channel are not aigned yet, the algorithm i repeated by remained channel and CPE If an available component carrier ha the r-th bet gain for everal CPE, the channel i allocated to the CPE with the larget gain and next iteration tart. The channel allocation procedure continue until a channel ha been allocated to N out of M CPE. If the number of elected CPE i not equal to N, the algorithm hould be applied for remained CPE and CC. After joint relay election and channel allocation procedure, the ret of the CPE that do not act a relay, communicate to each other a a econdary network (CR-CPE). A graphical interpretation of thi CPE diviion i hown in Fig. 5. In the next ection, we will explain how the power and channel have to be allocated to each of thee CPE, in order to maximize the econdary network throughput. Termination: The proce will be terminated if at leat one of the following condition i true: - The SINR experienced by UE ( γ ) goe below the threhold. - The tranmit power of the CBS goe above it CBS maximum tranmit power contraint ( P max ). A RE-CPE node can terminate the power updating proce by broadcating ome pecial tone (control meage). UE c IV. Power Updating for Cognitive Relay Network In thi ection, we conider the problem of power control to maximize the downlink throughput of the cognitive radio network while protecting primary receiver. We propoe a power updating mechanim that trie to maximize the coverage and throughput of the econdary network while guaranteeing the SINR contraint of the primary receiver. Note that the power updating proce i applied to one channel at a time. For the c-th channel, 1 c N, the following action are carried out: Initialization: CBS which operate on channel c, initiate the power updating proce by broadcating ome pecial tone. Therefore, the CBS et it tranmit power to the initial Fig. 5. A graphical interpretation of propoed network model. After applying joint relay election and channel allocation algorithm, the number of remain cognitive radio CPE (K) i 4 (M=8, N=4 and K=4) V. Channel Aignment by Bipartite Matching After the power updating proce for each channel c, the CBS ha a maximum tranmit power and CR-CPE have different SINR for the two tep of relaying. The problem i how to aign K channel to different CR- CPE o that the total downlink throughput i maximized during the relaying proce. Thi i achieved by firt tranforming the problem into a weighted bipartite matching and then, finding a maximal weighted match. We obtain a maximal weighted matching of a bipartite graph by the following procedure. Note that the um of Copyright 2011 Praie Worthy Prize S.r.l. - All right reerved Int. Journal on Communication Antenna and Propagation, Vol. 1, N

6 A. Alizadeh, S. M.-S. Sadough, S. A. Ghorahi the throughput in two tep i conidered a the weighted matching parameter. Maximal Weighted Bipartite Matching Procedure: Step 1: Start with empty match graph which edge are not elected. Step 2: Find a maximum augmenting path for the current match. The augmenting path i a path with edge alternating between matched and unmatched while the core of path i maximized. The core of an augmenting path i equal to the um of weight (rate) of unmatched edge ubtracted by the um of weight of matched edge. While the core of the maximum augmenting path i poitive, then go to Step 3. Otherwie, finih the procedure ince the current match i maximum. Step 3: Flip the maximum augmenting path obtained in Step 2. Unmatched edge of the path are changed to matched edge and the matched edge of the path are changed to unmatched one. Go back to Step 2 to find another maximum augmenting path and continue. Fig. 6. Link between CR-CPE and CBS a a bipartite graph; each edge how the SINR of the ith CPE (left edge) in the jth channel (right edge) By uing thi method, the total throughput of the econdary network i maximized and one channel i aigned to each CR-CPE. The path lo exponent i et to 3. The maximum order of relay election algorithm i et to 2. The noie power denity at each receiver i N 0 = 100dBm. The required SINR for UE i aumed to be between 6 and 20 db. The maximum tranmit power on each channel for CBS and are 60 dbm and initial value of power updating i 20 dbm. The caling factor ued in the power updating proce i δ = ~ 1 db. The central carrier frequency of the primary network i conidered 3500 MHz which i recommended by 3GPP in Rel.10. We aume that the tranmiion rate of the primary and econdary network can be conidered a a function of SINR. Thi function depend on variou factor uch a the available coding/modulation cheme and bit error rate requirement. According to [17], we ue the approximation f ( γ ) 13 / γ = for the rate function. 06. Our imulation are baed on a comparion between the propoed reource allocation cheme and imple model. In Fig. 7, we imulate the random relay election approach to compare with the performance of our propoed model. In thi Figure, only the relay election tage (ee Fig. 1) i contrated. Then, Fig. 7 how the effect of minimum required SINR of the primary receiver in the whole network. The throughput of the econdary network, imilarly, received SINR at CR-CPE i decreaed when the minimum SINR for the receiver of LTE-A ytem (UE) i increaed while our propoed relay election algorithm improve both parameter. Figure 8, 9 and 10 illutrate a comparion between our propoed algorithm and uniform power allocation in the econd tage of our reource allocation cheme. Fig. 8 how a comparion between the throughput achieved by uing our propoed algorithm and uniform power allocation cheme. VI. Simulation Reult We conider a quare ervice area equal to 1,000 m 1,000 m in which the ditance eparating and UE i equal to 1,000 m and a cognitive radio network i deployed between and UE. A CBS i conidered in the CR cell to erve a et of CR-CPE a well a all CPE. The total number of CPE i M = 15. The number of component carrier i et to 5. All CPE are randomly deployed acro the entire ervice area with a uniform ditribution. A ample network, with 8 CPE which four of them work a RE-CPE, i given in Fig. 5. Each component carrier i regarded a one channel in our joint power and channel allocation cheme. The fading channel i repreented by a Rayleigh ditribution channel, with three non-los path. Fig. 7. Throughput of primary and econdary network veru minimum SINR level at primary receiver (UE) for propoed cheme and random relay election approach Copyright 2011 Praie Worthy Prize S.r.l. - All right reerved Int. Journal on Communication Antenna and Propagation, Vol. 1, N

7 A. Alizadeh, S. M.-S. Sadough, S. A. Ghorahi power allocation cheme cannot adjut the interference impoed to the primary network and the SINR of the primary network will be le than that obtained with the power updating approach. Fig. 8. Comparion between the throughput achieved by uing our propoed model and thoe achieved by uing a uniform power allocation cheme veru the minimum required SINR at the primary receiver In the uniform power allocation cheme, there i no power updating algorithm. In the maximum uniform tranmit power allocation, the maximum tranmit power of CBS in each channel i aigned uniformly. Similarly, the initial tranmit power of CBS for power updating proce i aigned a minimum uniform power allocation. Fig. 9. Throughput of primary and econdary network veru the number of CPE (c=5) If the obtained tranmit power of econdary network doe not atify the minimum required SINR at the primary receiver, it i et to zero. A hown in Fig. 9, increaing the number of CPE affect on both the primary and econdary network and provide higher throughput. When the number of CPE increae, there are more choice for electing the CPE a CR-CPE. Therefore, the increaing of the CPE node can alo increae the throughput of the primary network. A comparion between the propoed algorithm and the maximum uniform power allocation cheme i illutrated in Fig. 10. Thi Figure how that the uniform Fig. 10. Throughput of the primary and econdary network veru the maximum tranmit power of the econdary network (CBS) VII. Concluion In thi paper, we propoed a reource allocation algorithm for LTE-A ytem which utilize the cognitive relaying cheme. We utilize carrier aggregation technology from LTE-A ytem which allow the multiple component carrier (CC) tranmiion over a frequency bandwidth up to 100 MHz. The reource allocation algorithm conit of three main tage. In the firt tage, the heuritic relay election algorithm wa propoed to aign higher gain cognitive relay link to the LTE-A network, and to divide the CPE into the relaying node and the cognitive radio network. After the relay election, the power updating method wa introduced to maximize the tranmit power of CR-CPE while keeping the SINR above a predefined threhold at the primary receiver. Thi procedure wa completed by utilizing the bipartite matching algorithm to allocate the bet component carrier to each CR-CPE and to maximize the total throughput of the CR network. Simulation reult howed that the throughput of the whole network i increaed by uing the propoed algorithm in comparion with conventional random relay election and uniform power allocation method. Reference [1] S. Haykin, Cognitive radio: brain-empowered wirele communication, IEEE Journal on Selected Area in Communication, vol. 23, pp , [2] S. Srinivaa and S.A. Jafar, The Throughput Potential of Cognitive Radio: A Theoretical Perpective, IEEE Communication Magazine, Vol. 45, Iue: 5, 2007, Page(): [3] A. Attar, S. A. Ghorahi, M. Sooriyabandara, A. H. Aghvami, Challenge of real-time econdary uage of pectrum, Computer Network (Elevier), Vol. 52, pp , Copyright 2011 Praie Worthy Prize S.r.l. - All right reerved Int. Journal on Communication Antenna and Propagation, Vol. 1, N

8 A. Alizadeh, S. M.-S. Sadough, S. A. Ghorahi [4] A. Hot-Maden and J. Zhang, Capacity bound and power allocation for wirele relay channel, IEEE Tranaction on Information Theory, vol. 51, no. 6, pp , June [5] S. Serbetli and A. Yener, Relay aited F/TDMA ad hoc network: Node claification, power allocation and relaying trategie, IEEE Tranaction on Communication, vol. 56, no. 6, pp , June [6] Y. S. V. Raman, B. Prabhakara Rao, S. Sri Gowri, Performance of Propoed Algorithm Baed on Ditributed Dynamic Channel Allocation, The International Journal on Communication Antenna and Propagation (IRECAP), Vol. 1 N. 1, pp , February [7] L. Elkahlan, M., Leung, C. and Schober, R. : Performance analyi of channel aware frequency hopping, IEEE Proceeding on Communication, Vol. 153, No. 6, pp , December [8] H. Luo, Z. Zhang, G. Yu, Cognitive Cooperative Relaying, 11th IEEE Singapore International Conference on Communication Sytem, [9] B. Zheng and Q. Meng, Optimal Power Allocation for Cognitive Relay Network, International Conference on Wirele Communication & Signal Proceing, [10] Q. Zhang, J. Jia, and J. Zhang, Cooperative Relay to Improve Diverity in Cognitive Radio Network, IEEE Communication Magazine, February [11] H. Chen, Relay Selection for Cooperative Spectrum Sening in Cognitive Radio Network, International Conference on Communication and Mobile Computing, [12] S. Akhavan Ataneh, S. Gazor, Reource Allocation and Relay Selection for Collaborative Communication, arxiv: v1, November 25, [13] A. Alizadeh, S. M. S. Sadough, and N. T. Khajavi, Optimal Beamforming in cognitive two-way relay netwrok, IEEE 21t International Sympoium on Peronal, Indoor and Mobile Radio Communication, PIMRC, 2010, pp [14] Y. Yang, H. Hu, J. Xu and G. Mao, Relay technologie for WiMax and LTE-advanced mobile ytem, IEEE Communication Magazine, October 2009 [15] K. Loa, C. Wu, S. Sheu and Yifei Yuan, IMT-advanced relay tandard, IEEE Communication Magazine, Augut [16] A. Saataki, K. Tagkari, D. von-hugo, M. Siebert, M. Roenberger, P. Demeticha, Cognitive Radio Reource Management for Improving the Efficiency of LTE Network Segment in the Wirele B3G World, 3rd IEEE Sympoium on New Frontier in Dynamic Spectrum Acce Network, DySPAN 2008, pp [17] L. G. U. Garcia, K. I. Pederen and P. E. Mogenen, Autonomou Component Carrier Selection: Interference Management in Local Area Environment for LTE-Advanced, IEEE Communication Magazine, September 2009, pp Author information Cognitive Telecommunication Reearch Group, Department of Electrical and Computer Engineering, Shahid Behehti Univerity G.C., Evin , Tehran, Iran. ar.alizadeh@mail.bu.ac.ir _adough@bu.ac.ir a_ghorahi@bu.ac.ir A. Alizadeh wa born in Raht, Iran in He received hi B.Sc. degree in Electrical Engineering (electronic) from Amirkabir Univerity of Technology, Tehran, Iran in He i currently puruing hi M.Sc. degree in Electrical Engineering (telecommunication) at Shahid Behehti Univerity, G. C., Tehran, Iran. Since May 2010, he ha been with the Cognitive Radio Reearch Group of Shahid Behehti Univerity, Tehran, Iran. Hi current reearch interet include cognitive radio, two-way relay network, reource allocation algorithm and convex optimization problem. S. M. S. Sadough wa born in Pari in He received hi B.Sc. degree in Electrical Engineering (electronic) from Shahid Behehti Univerity, Tehran, I.R. Iran in 2002 and the M.Sc. and hi Ph.D. degree in Electrical Engineering (telecommunication) from Pari- Sud 11 Univerity, Oray, France, in 2004 and 2008, repectively. From 2004 to 2007, he held a joint appointment with the National Engineering School in Advanced Technique (ENSTA), Pari, France, and the Laboratory of Signal and Sytem (LSS), at Supélec, Gif-ur-Yvette, France. He wa a lecturer in the Department of Electronic and Computer Engineering (UEI), ENSTA, where hi reearch activitie were focued on improved reception cheme for ultra-wideband communication ytem. From December 2007 to September 2008, he wa a potdoctoral reearcher with the LSS, Supélec-CNRS, where he wa involved in the European reearch project TVMSL with Alcatel-Lucent France. Since October 2008, he ha been a member of the Faculty of Electrical \& Computer Engineering, Shahid Behehti Univerity, where he i currently an Aitant Profeor in the Department of Telecommunication. Dr. Sadough' area of reearch include ignal proceing, communication theory, and digital communication. S. A. Ghorahi received hi B.Sc. and M.Sc. degree in Electrical Eng. from the Univerity of Tehran, Iran, in 1992 and 1995, repectively. Then, he joined SANA Pro Inc., where he worked on modelling and imulation of OFDM baed wirele LAN ytem and interference cancellation method in W-CDMA ytem. Since 2000, he worked a a reearch aociate at King College London on capacity enhancement method in multilayer W-CDMA ytem ponored by Mobile VCE. In 2003 He received hi PhD at King College and ince then he worked at King College a a reearch fellow. In 2006 he joined Samung Electronic (UK) Ltd a a enior reearcher and now he i a faculty member of Cognitive Telecommunication Reearch Group, Department of Electrical Engineering, Shahid Behehti Univerity G.C., at Tehran, Iran, working on wirele communication. Copyright 2011 Praie Worthy Prize S.r.l. - All right reerved Int. Journal on Communication Antenna and Propagation, Vol. 1, N