Enhancing IEEE n WLANs using group-orthogonal code-division multiplex

Transcription

1 Teecommun Syst (2008) 38: DOI /s x Enhancing IEEE 80211n WLANs using group-orthogona code-division mutipex Guiem Femenias Feip Riera-Paou Pubished onine: 18 March 2008 Springer Science+Business Media, LLC 2008 Abstract The definition of the next generation of wireess networks is we under way within the IEEE High Throughput Task Group committee The resuting standard, to be caed IEEE 80211n, is expected to be a backwardcompatibe evoution of the successfu IEEE 80211a/g systems aso based on muticarrier techniques It can be anticipated that 80211n systems wi outperform its predecessors in terms of transmission rate and/or performance, mainy, due to the use of mutipe antennae technoogy for transmission and reception In this paper we propose to incorporate group-orthogona (GO) code division mutipex (CDM) into the IEEE 80211n specifications to further enhance its performance It is shown how GO-CDM can take fu advantage of the diversity offered by the mutipe antennae and muticarrier transmission by using an iterative maximum ikeihood (ML) joint detector Furthermore, the use of GO-CDM does not compromise the backward compatibiity with egacy systems Keywords IEEE 80211n Muticarrier Code-division mutipex Maximum ikeihood detection Turbo receivers 1 Introduction The ast decade has seen an exposive growth in the depoyment of wireess oca area networks (WLANs) which G Femenias F Riera-Paou ( ) Mobie Communications Group, Dept of Mathematics and Informatics, University of the Baearic Isands, Pama, Maorca (Ies Baears), Spain e-mai: feipriera@uibes G Femenias e-mai: guiemfemenias@uibes has made the concept of nomadic computing a reaity Most of these networks are based on one of the favors of the IEEE famiy of standards The most recent and powerfu set of specifications currenty depoyed, IEEE 80211g, operates on a bandwidth of 20 MHz and its physica ayer is based on a particuar form of muticarrier transmission, namey, orthogona frequency division mutipexing (OFDM) enabing these systems to achieve transmission rates up to 54 Mbps At present, the standardization of what shoud be the next step, named IEEE 80211n, is being pursued by the IEEE High Throughput Task Group committee The new standard wi support much higher transmission rates thanks to the use of mutipe antenna technoogy and other enhancements such as the possibiity of operating on a 40 MHz bandwidth (empoying more subcarriers) and transmission modes using a reduced guard interva In its fastest mode, 80211n is expected to surpass a transmission rate of 700 Mbps Despite a the enhancements introduced, it is mandatory for the new standard to remain compatibe with muticarrier egacy systems (80211a/g) and therefore, 80211n-compiant devices shoud have means to fa back to oder specifications Like its predecessors, 80211a and 80211g, the new 80211n standard wi aso be based on OFDM aowing the use of other enhancements which have recenty been proposed for this type of physica ayers A powerfu improvement over conventiona OFDM was the introduction of muticarrier code division mutipex (MC-CDM) OFDM by Kaiser in [1] In MC-CDM, rather than transmitting a singe symbo on each subcarrier as in conventiona OFDM, groups of symbos are mutipexed together by means of orthogona spreading codes and simutaneousy transmitted on a group of subcarriers This technique resembes very much the principe behind muticarrier code-division mutipe access (MC-CDMA) [2] where different users share a group

2 38 G Femenias, F Riera-Paou of subcarriers by using each of them a different spreading code More recenty, group-orthogona MC-CDMA (GO- MC-CDMA) [3, 4] has been introduced as a particuar favor of MC-CDMA whereby users are spit in groups and each group excusivey uses a (sma) subset of a the avaiabe subcarriers The subcarriers forming a group are chosen to be as separate as possibe in the avaiabe bandwidth in order to maximize the frequency diversity gain A GO- MC-CDMA setup can be seen as many independent MC- CDMA systems of ower dimension operating in parae This reduced dimension aows the use of optimum receivers for each group based on maximum ikeihood detection at a reasonabe computationa cost Group-orthogonaity has aso been proposed for (uncoded) MC-CDM systems in [5] where resuts are given for group dimensioning and spreading code seection Nevertheess, as shown in [6], the benefits of CDM-OFDM are rather imited when measuring the coded performance in typica operating scenarios conforming to IEEE 80211a specifications, especiay, when using decoders based on soft/iterative procedures This is due to the arge subcarrier correation found in many wireess environments which severey imits the achievabe frequency diversity This issue is argey soved with the introduction of mutipe transmit and receive antennae as then subcarrier correation within a group can be greaty reduced by expoiting the spatia dimension A comparison between two of the possibe techniques to expoit the spatia diversity, namey, space-time bock coding (STBC) and cycic deay diversity (CDD), within the context of IEEE 80211n has recenty appeared in [7] but in there, the receiver structures proposed are based on inear techniques (minimum mean square error detectors) and hence are not suitabe for iterative processing This paper, after reviewing the physica ayer of the current IEEE 80211n draft, proposes the use of GO-MC-CDM to further improve its performance It wi be shown how using an iterative reception scheme based on a maximum ikeihood mutisymbo detector (ML-MSD) with soft decoding aows the expoitation of both, frequency and spatia diversity Simuation resuts are presented for two possibe spatia configurations, one uniquey based on cycic deay diversity (CDD) and another one based on the combination of spatia division mutipexing (SDM) and CDD We concude this introduction with a brief notationa remark: throughout this paper vectors and matrices are denoted with ower and upper cases bod characters, respectivey Scaars are represented with non-bod characters (either ower or upper case) We use the notation D(x) to denote a diagona matrix with vector x at its main diagona Vectors are assumed to be coumn-oriented and the notation ( ) T is used to denote the transposition of a vector or matrix Finay, x [k] represents a vector x with its kth entry removed 2 System mode for current IEEE 80211n proposa The current draft for the physica ayer of IEEE 80211n can be found in [8] As mentioned in the introduction, a quaitative difference of IEEE 80211n with respect to previous standards is the introduction of mutipe antenna technoogy Moreover, the standard is defined in such a way that aows the use of different methods to expoit the spatia dimension, namey, spatia division mutipexing (SDM), cycic deay diversity (CDD) and space-time bock coding (STBC) In this work we have focused on the design of an enhanced physica ayer based on the appication of GO-CDM when using CDD and/or SDM 1 The transmission architecture is depicted in Fig 1 Note that this figure aready incudes the optiona GO-CDM extension (in dashed ines) which wi be covered in detai in the next section We focus on the IEEE 80211n specification operating in a bandwidth of B = 20 MHz utiized by means of N c = 64 orthogona subcarriers of which Nc d = 52 subcarriers are used to transmit user data whie the rest, Nc p = 12, correspond to piot and guard subcarriers Transmitter and receiver are assumed to have N t and N r antennas, respectivey, with 1 N t, N r 4 The user data to be transmitted is generated and segmented into frames of N b bits It is assumed that frames are independent of one another, and therefore, to simpify notation, there is no need to specify the segment index Each generated frame is subsequenty fed to a rate R c punctured convoutiona encoder (RCPCC box in Fig 1) to generate N b /R c coded bits The coded bits are then separated into N s {1, 2, 3, 4} spatia streams as specified by the spatia parsing equation in [8] It shoud be pointed out that for the cases of empoying CDD and/or SDM, it shoud hod that N s N t The bits on each stream are then intereaved (bocks Π 1 in Fig 1) and subsequenty mapped to symbos from an m-point consteation (BPSK, m-qam) with m = 2 M where M represents the number of bits per symbo, yieding the set of signa vectors to be transmitted s w = (s1 w sw 2 sw N qam ) T where 1 w N s represents the spatia stream index and N qam = N b /(R c N s M) denotes the number of symbos per frame to be transmitted on each spatia stream The symbos on the different spatia streams are at this point propery segmented into N OFDM OFDM symbos, each made of N d (QAM) symbos hoding that N qam = Nc d N OFDM We wi denote by s w,1 N OFDM, the th OFDM symbo of the wth spatia stream in the current frame Skipping for the moment, the GO-CDM processing, the resuting OFDM symbos from a streams are ineary mixed according to (x u )T = W (s 1 s 2 s N s ) T (1) 1 The appication of GO-CDM in combination with STBC in the context of IEEE 80211n is a topic of current research

3 Enhancing IEEE 80211n WLANs using group-orthogona code-division mutipex 39 Fig 1 IEEE 80211n transmitter proposa supporting SDM and/or CDD where 1 u N t and W represents an N t N s spatia spreading matrix The roe of the matrix W is twofod: on one hand, for the cases where N s <N t, the spatia spreading serves to distribute the incoming streams among a transmit antennae thus making fu use of the avaiabe spatia diversity On the other hand, the specific seection of W, in combination with a suitabe choice of the cycic deays δ 1,,δ Nt (see Fig 1), provides the system with either CDD, SDM or a combination of the two Exampes are ater provided showing how this matrix and cycic deays can be chosen to expoit the spatia diversity in a prescribed manner Finishing the transmission procedure, and as shown in Fig 1, the different spatiay spread OFDM symbos, x u, are then IFFT-converted, expanded with a cycic prefix and processed by the RF transmission chains Taking for granted that transmit and receive antennas are sufficienty apart, the N t N r propagation channes between transmitter and receiver are safey assumed to be independent and derived from a common and scenario-dependent power deay profie P 1 S(τ) = φ(p)δ(τ τ p ) (2) p=0 where P denotes the number of independent paths of the profie and φ(p) and τ p denote the power and deay of each path, respectivey A singe reaization of the channe impuse response between transmit antenna a t and receive antenna a r at time instant t wi then have the form P 1 h a t,a r (t; τ)= h a t,a r p (t)δ(τ τ p ) (3) p=0 where it wi hod that E[ h a t,a r p (t) 2 ]=φ(p) The corresponding frequency response wi be given by P 1 h a t,a r (t; f)= h a t,a r p (t)exp( j2πf τ p ) (4) p=0 which, evauated at the N c equispaced subcarrier frequencies, f 0,f 1,,f Nc 1, yieds the N c 1 vector h a t,a r (t) = ( h a t,a r (t; f 0 ) h a t,a r (t; f Nc 1)) T Assuming without oss of generaity that the channe is static over the duration of a frame and independent from other frames, to simpify the notation, we wi express the frequency response of the data subcarriers (ie piots and guard subcarriers excuded) simpy as h a t,a r = ( h a t,a r 0 h a t,a r 1 h a t,a r Nc d 1)T At this point it is necessary to distinguish which form of spatia diversity is empoyed, which is basicay determined by the choice of W and cycic deays In setups based excusivey on SDM, it wi hod that N s = N t and δ at = 0 for 1 a t N t This impies that each antenna is in charge of transmitting one spatia stream In contrast, in configurations where transmit antennae are configured to expoit CDD, it wi hod that N s = N t /Λ, where Λ denotes the number of antennae invoved in the CDD processing (CDD factor) of each spatia stream That is, when using CDD, the avaiabe antennae are spit in N s groups of Λ antennae and each group is used for the transmission of one spatia stream In CDD, a the antennae in a CDD group transmit the same information but each antenna appies a different cycic deay [9] It is easy to show [9] that this amounts to transmit a given spatia stream information from a singe antenna over the composite channe reaching antenna a r with frequency response h a r w = ( h a r 0,w h a r Nc d 1,w)T where Λ 1 h a r q,w = a w t =0 h aw t,a r q ( exp j 2πqδ at w N c with h aw t,a r q denoting the different channe frequency responses on the qth subcarrier between the antennae beonging to CDD group w and receiver antenna a r and δ a w t represents the specific cycic deay appied by each antenna in the CDD group Taking into account that the avaiabe transmit antennae wi be spit into parae CDD systems, it can be derived from [9] that an advantageous deay choice wi be given by { 0, if mod(at 1,Λ)= 0 δ at = (6) N c /(N t /Λ) + δ at 1, otherwise where mod(a, b) denotes the moduus of a/b It is important to stress that, to a effects, the composite CDD channe h a r w can be considered as a conventiona channe with the distinctive feature of an increased frequency seectivity with respect to the origina power profie defined by S(τ) Having defined the transmission scheme and channe, it is now possibe to specify the reception equation for the N s ) (5)

4 40 G Femenias, F Riera-Paou OFDM-symbos which have been transmitted in parae at a given instant as r 1 H 1,1 H 1,N s x 1 = + r N r H N r,1 H 1,N s x N s υ 1 υ N r, where H i,j wi be of the form D( h a r,a t ) for the case of SDM (N t = N s ) or the form D( h a r w ) for the case where CDD is used (N s = N t /Λ) The Nc d 1 vectors υ1 g,,,υn r g, represent the additive white Gaussian noise sampes on each receiving antenna Every scaar noise sampe, is assumed to be distributed according to a zero-mean Gaussian pdf with variance σn 2 The origina data symbos to be transmitted sw i, are suitaby scaed to have power E{ si w 2 }=1inthe case of SDM transmission and E{ si w 2 }=1/Λ if the CDD component is present In this atter case, this factor represents the distribution of energy avaiabe per symbo among the various transmit antennas used for the transmission of the wth spatia stream For both cases, SDM and CDD (and their combinations), this power normaization aows the operating signa to noise ratio to be written as E s /N 0 = 1/σn 2 3 Group-orthogona CDM for IEEE 80211n The introduction of mutipe transmit and receive antenna brings aong the possibiity of further expoiting the potentia frequency seectivity, speciay, whenever a CDD component is present To this end, GO-CDM can be used to distribute the source symbos energy among mutipe subcarriers Figure 2 shows a detaied diagram of the GO-CDM mechanism which woud correspond to the dashed boxes in Fig 1 The GO-CDM processing is performed on each spatia stream and consists of spitting the incoming OFDM symbos, s w,intog groups of Q = Nc d /G QAM symbos each Each group, s w g,, is then assigned Q subcarriers which wi be used to jointy transmit the Q source symbos in the group The distribution of the group symbos among the Q subcarriers is carried out by means of a spread and mutipex Fig 2 GO-CDM extension (7) operation defined by ŝ w g, = Csw g, (8) with 1 g G and with C = (c 0 c 1 c Q 1 ) representing the Q Q spreading matrix with each c q denoting the Q 1 spreading code associated with the q th symbo in the g th group A typica choice of spreading matrix is the Wash- Hadamard matrix due to its coumn-wise orthogonaity and ow compexity impementation We note that the spreading matrix is common to a groups since each group utiizes separate (and orthogona) sets of subcarriers and therefore, there is no need to provide extra separation The resuting spread and mutipexed symbos from a groups are then time intereaved (bock Π 2 in Fig 2) to increase the resiience against noise The idea is to transmit (QAM spread) symbos from different OFDM-symbos in each group so that, upon deintereaving at the receiver, each symbo in the group has experienced uncorreated noise sampes To avoid confusion with the spatia spreading performed by the inear transformation given by W, from this point onwards, we wi refer to the GO-CDM processing as tempora spreading As can be inferred from Fig 1, if GO-CDM processing is present, the spatia spreading in (1) operates on the temporay spread symbos resuting from (8) rather than on s 1,,sN s We concude this section by noting that backward compatibiity with IEEE 80211a/g can easiy be insured by simpy setting the group size to Q = 1 4 ML-based turbo reception As in conventiona OFDM systems, the reception process on each receive branch (iustrated in Fig 3) begins by removing the cycic prefix (CPR) and performing the FFT on each of the OFDM-symbos forming the transmitted frame In case GO-CDM has been empoyed at the transmitter, as we assume through the rest of this section, the corresponding deintereaving (Π 1 2 ) is then appied resuting in N r vectors r a r of dimension Nc d 1 (piots and guard subcarriers are not taken into account) These vectors are spit into groups, r g,, in concordance with the grouping performed at transmission It shoud be noted now that, as inferred from (7), each received group vector r g, contains information from the groups of a transmit spatia streams s w g,, making mandatory some form of interference canceing device to reduce the intersymbo interference in order to estimate the different symbo groups from the different spatia streams The main reason to spit the symbos in groups when spreading and mutipexing (ie appying group-orthogonaity) is to make ML-based detection computationay feasibe within each singe group In ine with this, we have based our detection mechanism on a bank of ist sphere detectors (LSD)

5 Enhancing IEEE 80211n WLANs using group-orthogona code-division mutipex 41 Fig 3 ML-based turbo receiver for GO-CDM enhanced IEEE 80211n [10], each targeting a singe-group The LSD detector is based on the sphere detection procedure introduced in [11] which is an efficient method of performing an exhaustive search among a set of candidates (ie ML detection) The main feature of LSD is that, apart from symbo estimates, it is abe to produce soft information regarding the bits of the estimated symbo in the form of ikeihood ratios (LLRs) Since the LSD detector works on a group basis, it is usefu at this point to write the reception equation for a singe group as r 1 g, r N r g, = Hg 1,1 W C H 1,N s g W C H N r,1 g W C H 1,N s g W C υ 1 g, + υ N r g, or, more compacty, s 1 g, s N s g, (9) r g, = A g s g, + υ g, (10) where r g, = (r 1 g, r N r g, )T represents the N r reception vectors evauated on the frequencies assigned to the gth group, s g, = (s 1 g, s N s g, )T are the symbos from a transmitted spatia streams in group g, A g corresponds to the matrix in (9) jointy modeing the channe, tempora spreading and spatia spreading effects, and υ g, = (υ 1 g, υ N r g, )T are the noise sampes added on the group subcarriers We define now the mapping s g, = M(b) as the moduation mapping to arrive to symbo vector s g, comprising a the symbos beonging to group g of the th OFDM symbo from the N s transmitted spatia streams from the corresponding group bits b = (b 1 b 2 b (Ns MQ)) T (to simpify notation, we skip the group and OFDM-symbo indices when referring to the bits) Making use of the Max-og approximation, the LLR for a given bit, b i, can then be approximated by [10] L D1 g, (b i) 1 { 2 max 1 b B i,+1 σn 2 r g, A g M(b) 2 } + b T [i] (LA1 g, ) [i] 1 { 2 max 1 b B i, 1 σn 2 } + b T [i] (LA1 g, ) [i] r g, A g M(b) 2 (11) where the symbos B i,+1 and B i, 1 represent the sets of 2 NsMQ 1 bit vectors whose i th position is a +1 or 1, respectivey The (N s MQ 1) 1 vector (L A1 g, ) [i] contains the a-priori LLR for each bit in b except for the i th bit A a-priori LLRs are assumed to be zero for the first iteration Since moderate vaues of M and/or Q make the sets B i,+1 and B i, 1 extremey arge and therefore the search in (11) computationay unfeasibe, LSD imits the search to the sets ˆB i,+1 = B i,+1 C and ˆB i, 1 = B i, 1 C where C is the set containing the bit vectors corresponding to the N cand group candidates coser, in an Eucidean sense, to the received group vector, ie, C ={b 1,,b N cand } where b j = M 1 ( s j g, ) with { s1 g,,, sn cand g, } being the N cand group candidates for which r g, A g s g, 2 (12) is smaest Notice that when computing the ist of group candidates using (12), the a-priori information from each bit is not taken into account However, when N cand is chosen sufficienty arge most often the ist of candidates minimizing (12) contains with high probabiity the maximizer (ML soution) of (11) Two points are to be taken into account when appying the LSD Firsty, it is possibe that either ˆB i,+1 or ˆB i, 1 is empty for a given bit Such a situation arises when for a given bit position, a the candidates turn out to have a +1 (or 1 ) in that position In these cases,

6 42 G Femenias, F Riera-Paou that bit position is assigned an extreme (positive or negative) vaue [10] Secondy, the imitation of the space search to a imited poo of candidates may cause the a-priori information {L A1,1,,L A1,N s } to be of imited vaue or even deeterious whenever the candidate poo does not contain the soution hinted by the a-priori information This probem can be aeviated using the constrained LSD detector proposed in [12] The computed LLRs for the bits in the G groups of the N OFDM OFDM-symbos for the N s spatia streams, L D1 g, = (LD1 g, (1) LD1 g, (N smq)) are then de-grouped and de-segmented to arrange the LLRs in frame format for the N s transmitted streams {L D1,1,,L D1,N s } The a-priori information L A1 is subtracted from the computed LLRs and the resuting extrinsic information {L E1,1,,L E1,N s } is then temporay de-intereaved and spatiay merged (eg inverse spatia parsing) resuting in the signa abeed as L A2 in Fig 3, which is fed to a soft-in/soft-out maximum a- posteriori (MAP) decoder The MAP decoder output can be suppied to a bit sicer to obtain bit estimates or fed back to the LSD modue for further LLR refinement In this atter case, see Fig 3, ony extrinsic information (L E2 )issent back to the detector 5 Numerica resuts Numerica simuations have been conducted to evauate the performance of the GO-CDM extension in a IEEE 80211n setup The system has been configured to operate on a 20 MHz bandwidth with a tota of N c = 64 subcarriers of which Nc d = 52 are used to carry data Without oss of generaity, information bits are generated in packets of N b = 416 bits The transmitter uses a rate terminated punctured convoutiona coder with generator poynomias [ ] Perfect channe knowedge and subcarrier synchronization are assumed at the receiver It has been aso presupposed that the duration of the cycic prefix appended to each OFDM symbo exceeds the root mean squared channe (rms) deay spread, thus eiminating any interference among consecutive OFDM symbos As in [7], we have configured the system to have N t = 4 and N r = 2 transmit and receive antennae, respectivey The simuations have been conducted using channe mode E-NLOS as defined in [13] This channe profie corresponds to a arge office environment and it is made of 38 independent paths distributed among 4 custers and it has an rms deay spread of 100 ns Two different spatia configurations have been tested by considering different numbers of spatia streams and particuar seections of the spatia spreading matrix In a simuations, an ML-based turbo reception scheme (with 2 iterations) as the one introduced in the previous section was empoyed with the number of considered candidates set to N cand = 64 For comparison, the resuts obtained with hard and soft non-iterative Fig 4 SDM-CDD transmission N t = 4, N r = 2, N s = 2 MCS 9 Viterbi decoding wi aso be shown In the first spatia configuration, the four antennas are configured as two CDD systems (CDD factor Λ = 2) operating in parae It is easy to see that this setup can be achieved with a choice of spatia spreading matrix given by: 1 0 W = (13) 0 1 and cycic deays defined by δ 1 = δ 3 = 0 and δ 2 = δ 4 = 32 This system is an exampe of a combination of CDD and SDM, as two independent spatia streams are simutaneousy transmitted over different antennae and each stream empoys CDD The system is configured to use QPSK moduation which, in combination with the chosen number of spatia streams and bandwidth, woud correspond to the moduation coding scheme (MCS) number 9 in [8] achieving a transmission rate of 26 Mbps Figure 4 shows the attained packet error rate (PER) when empoying GO-CDM processing with Q = 4 (back ines with soid markers) in comparison with the standard system without GO-CDM (Q = 1, red ines with hoow markers) It can be seen that the configuration with GO-CDM ceary outperforms the conventiona system when both empoy iterative detection In particuar, gains between 1 and 2 db are observed across the range of PER vaues of practica interest Note that in the standard configuration, the iterative detection hardy provides any PER reduction over soft Viterbi decoding In the second tested configuration ony one spatia stream is assumed to be present (N s = 1) and the four transmit antennae are used to perform CDD (Λ = 4) Choosing W = (1 111) T and setting the cycic deays to δ 1 = 0, δ 2 = 16, δ 3 = 32 and δ 4 = 48, configures the system as a 4- branch CDD setup For this scenario, the moduation format

7 Enhancing IEEE 80211n WLANs using group-orthogona code-division mutipex 43 Fig 5 CDD transmission N t = 4, N r = 2, N s = 1 MCS 3 has been set to 16-QAM (MCS number 3 in [8]) resuting again in a transmission bit rate of 26 Mbps The resuts with (Q = 4) and without (Q = 1)GO-CDMareshowninFig5 As in the previous spatia configuration, the simuations outcome eaves no doubt that the system with GO-CDM has the potentia, using iterative detection, to significanty outperform the current standard 4 Riera-Paou, F, Femenias, G, & Ramis, J (2006) On the design of group-orthogona MC-CDMA systems In Proceedings IEEE SPAWC, Cannes, France, Juy Riera-Paou, F, Femenias, G, & Ramis, J (2006) Downink performance of group-orthogona muticarrier systems In Proceedings IFIP PWC (pp ), Abacete, Spain, September Riera-Paou, F, & Femenias, G (2007) Combining muticarrier code-division mutipex with cycic deay diversity for future WLANs In Proceedings IEEE SPAWC, Hesinki, Finand, June Yan, W, Sun, S, Li, Y, & Liang, Y (2006) Transmit diversity schemes for MIMO-OFDM based wireess LAN systems In Proceedings IEEE PIMRC (pp 1 5), Hesinki, Finand, September Mujtaba, S A (2005) TGn sync proposa technica specification doc:ieee /0889r7, Draft proposa, Juy Bauch, G, & Maik, J (2006) Cycic deay diversity with bitintereaved coded moduation in orthogona frequency division mutipe access IEEE Transactions on Wireess Communications, 8, Hochwad, B, & ten Brink, S (2003) Achieving near-capacity on a mutipe-antenna channe IEEE Transactions on Communications, 51, Fincke, U, & Pohst, M (1985) Improved methods for cacuating vectors of short ength in a attice, incuding a compexity anaysis Mathematica Computing, 44, Liu, J, & Li, J (2005) Turbo processing for an OFDM-based MIMO system IEEE Transactions on Wireess Communications, 4, Erceg, V (2003) Indoor MIMO WLAN channe modes doc: IEEE /871r0, Draft proposa, November Concusions This paper has proposed the introduction of grouporthogona code-division mutipex within the context of the IEEE 80211n (draft) standard A reception scheme has been deveoped using an iterative ML-based detection procedure Simuations resuts have shown that the combination of GO- CDM with iterative detection brings aong important performance gains in terms of PER reduction irrespective of the spatia configuration used Moreover, the GO-CDM extension is simpe to incorporate to the current standard draft and it can easiy be configured to be backward compatibe with IEEE 80211a/g Acknowedgements This work has been supported in part by the MEC and FEDER under project MARIMBA (TEC ), Govern de es Ies Baears under project XISPES and grant PCTIB- 2005GC1-09, and a Ramon y Caja feowship, Spain References 1 Kaiser, S (2002) OFDM code-division mutipexing in fading channes IEEE Transactions on Communications, 50, Yee, N, Linnartz, J-P, & Fettweis, G (1993) Muti-carrier CDMA in indoor wireess radio networks In IEEE PIMRC (pp ), Yokohama, Japan, September Cai, X, Zhou, S, & Giannakis, G (2004) Group-orthogona muticarrier CDMA IEEE Transactions on Communications, 52(1), Guiem Femenias was born in Petra, Spain, in 1963 He received both the Teecommunication Engineer degree and the PhD degree in Teecommunications from the Universitat Poitècnica de Cataunya (UPC), Spain, in 1987 and 1991, respectivey From 1987 to 1994, he was a Researcher at UPC, where he became an Associate professor in 1990 Since 1995 he has been in an Associate Professor position at the Department of Mathematics and Informatics of the Universitat de es Ies Baears (UIB), Spain His current research interests and activities span the fieds of digita communications theory and wireess persona communication systems, with particuar emphasis on MIMO cross-ayer design in radio resource management strategies appied to fourth generation systems Dr Femenias has been the Project Manager of projects ARAMIS, DREAMS, DARWIN and MARIMBA, funded by the Spanish and Baearic Isands governments In the past, he was aso invoved in some European projects (ATDMA, CODIT, COST) Dr Femenias was a corecipient of the Best Paper Award at the IFIP Internationa Conference on Persona Wireess Communications 2007 Feip Riera-Paou wasborninpama(maorca, Spain) in 1973 He received the MS degree in Computer Engineering from the University of the Baearic Isands (UIB), (Maorca, Spain) in 1997, the MSc and PhD degrees in Communication Engineering from the University of Bradford (United Kingdom) in 1998 and 2002, respectivey, and the MSc degree in Statistics from the University of Sheffied (United Kingdom) in 2006

8 44 G Femenias, F Riera-Paou From May 2002 to March 2005, he was with Phiips Research Laboratories Eindhoven (The Netherands) first as a postdoctora feow (Marie Curie program, European Union) and ater as a member of technica staff At Phiips he was invoved in research programs reated to broadband audio/speech compression and speech enhancement for mobie handsets In Apri 2005 he became a research associate (Ramon y Caja program, Spanish Ministry of Science and Education) in the Mobie Communications Group of the Dept of Mathematics and Informatics at UIB where his work focuses on signa processing techniques for future wireess communication systems