TH`ESE - Library of Ph.D. Theses | EURASIP

XXII SIMPÓSIO BRASILEIRO DE TELECOMUNICAÇÕES - SBrT’05, 04-08 DE SETEMBRO DE 2005, CAMPINAS, SPA new Criterion for Transmit Diversity andBeamforming in Mobile CommunicationsDanilo Zanatta Filho and Luc FétyAbstract— We consider a wireless communication system inwhich the base station is equipped with an antenna array and themobile user has a single antenna. Classically, in this configuration,the antenna array at the base station is used to perform eitherbeamforming or transmit diversity. We propose a scheme thatexploits the antenna array to perform beamforming in order tomaximize the diversity of the channel seen at the receive end,i.e., the mobile user. This scheme is based on the criterion ofminimizing the variation of the power received by the mobileuser due to fading. Moreover, we propose an adaptive method tooptimize this criterion, which is a type of CM (Constant Modulus)criterion.Keywords— Adaptive antennas, smart antennas, downlinkbeamforming, transmit beamforming, transmit diversity, downlinkdiversity, CM, Constant Modulus, CMA, Constant ModulusAlgorithm, mobile wireless systems.I. INTRODUCTIONThe evolution of the wireless communication systems towards3GB (3 rd Generation and beyond) is propelled byInternet access and increasing demand for data-based services.Most of these services, such as internet surfing, are downlinkintensive,in opposition to voice services, which demand thesame data rate in both directions. These factors lead to anincreasing demand for higher system capacity and higher datarates, mostly in the downlink, which can be achieved by abetter link quality in terms of higher decoding SNR (Signalto Noise Ratio).The wireless propagation channel is formed by multiplepaths that sum with each other to construct the channel. Thissum of multiple paths leads to a fluctuation of the receivedpower in time, which is known as fading. Poor performancedue to prolonged deep fading of the channel is one of the mainproblems faced when using the wireless channel.In order to counteract deep fading, one can benefits fromthe decorrelation of the channel to combine uncorrelatedcopies of the same signal. It is very unlikely that all thesecopies are in a fade simultaneously. Hence, the chance thata deep fading occurs is greatly reduced and, when it occurs,its duration is also reduced when compared to the originalscenario of only one copy of the signal. This strategy is calleddiversity and it may be the single most important contributorto reliable wireless communications. Increasing the numberof uncorrelated copies of the same signal has the effect ofreducing the probability that the combined signal is in a fade.Danilo Zanatta Filho is with the DSPCom-DECOM-FEEC, UniversidadeEstadual de Campinas (UNICAMP), Brazil and the Laboratoire d’Électroniqueet Communication, Conservatoire National des Arts et Métiers (CNAM), Paris,France. Luc Féty is with the Laboratoire d’Électronique et Communication,Conservatoire National des Arts et Métiers (CNAM), Paris, France. E-mails:daniloz@decom.fee.unicamp.br and fety@cnam.frThe number of possible copies is called the diversity order ofthe channel.It is well established that the use of multiple antennas canimprove the performance of a wireless communication systemin a fading environment [1], [2]. From the point of view ofdiversity, the use of multiple antennas increases the diversityorder due to spatial uncorrelated copies of the signal, leadingto a diversity order greater than one even for flat fadingchannels. Although multiple antennas may be employed ateither the base station (BS), mobile unit (MU), or both, it ismost cost effective and practical to employ multiple antennasat the base station only. Hence, this article is restricted to thecase of employing multiple antennas at the BS.The considered case of employing multiple antennas atthe BS only has already been contemplated in the literaturefor receive diversity (RD) in the uplink [2]. The idea is tocombine the antennas outputs according to some criterion inorder to take advantage of the spatial diversity of the receivedsignal. This is called receive diversity since the diversity isexploited at the receiver. This is a well studied problem in theliterature and the main drawbacks to practical implementationare the required computational cost and, mostly, hardwareimplementation issues.However, the more challenging problem of profiting fromthe diversity in the downlink is yet an open problem. Recentworks propose to use some processing prior to transmissionby the multiple antennas in order to achieve a diversity gainat the MU receiver [3]. Since the processing is done at thetransmission side, this scheme is called transmit diversity(TD). With the use of TD, the MU receiver is kept simple (onlyone antenna) and yet profits from diversity gain. The maintheoretical impairment for TD is that the channel is unknownat transmission time.Among the different schemes for TD proposed in the literature,we can identify two main categories: one based on theknowledge of the downlink channel and another which doesnot require the knowledge of the downlink channel. The mainidea behind the TD scheme without channel knowledge is thateach antenna transmit an uncorrelated version of the desiredsignal. The receiver (MU) can hence combine these differentversions of the desired signal in order to profit from the spatialdiversity of the multipath channel. Since each antenna is usedregardless of the others, the radiation is omnidirectional andit is not possible to perform beamforming. In a multiusercontext, this approach has the weakness of generating a highinterference level due to omnidirectional radiation. Examplesof these schemes, called open loop TD, are the delay diversityschemes [4], [5] and space-time codes [6], with the Alamoutischeme [7] being one particular case.