OnCarrierFrequencyOffsetsinAlamouti-codedOFDMsystems similar to IEEE 802.11aLars Brötje, Sven Vogeler, and Karl-Dirk KammeyerUniversity of Bremen, Department of Communications Engineering,P.O. Box 33 04 40, D-28334 Bremen, Germany,E-mail: broetje, kammeyer¡ vogeler, @ant.uni-bremen.deandReinhard Rückriem and Stefan FechtelInfineon Technologies AG,P.O. Box 80 09 49, D-81609 Munich, GermanyE-mail:stefan.fechtel¡ reinhard.rueckriem, @infineon.comAbstract— In the presented paper, the influence ofCarrierFrequencyOffsets (CFO) inOFDMsystemswith transmit diversity using the Alamouti codingscheme is investigated. The OFDM system parametersare choosen according to the WLAN standardIEEE 802.11a. Different methods for estimating theCFO based on a new, only slightly modificated IEEE802.11a preamble are shown and the performance ofthese algorithms is compared to the single antennacase (SISO system). AlamouticodedOFDMsystemsare significantly more sensitive to carrier frequencyoffsets as we will show in our simulation results aswell as in measurements in the 2.4 GHz ISM band.Index Terms— OFDM, IEEE 802.11a, Alamouti,transmit diversity, carrier frequency offsetI. INTRODUCTIONMultiple Input Multiple Output (MIMO) systemsare well known for increasing system capacity comparedwith Single Input Single Output (SISO) systems.If there is no possibility to use multiple receiveantennas, one have to focus on transmit diversityschemes like space time block codes. Perhaps themost popular space time block code for two transmitantennas is the Alamouti scheme, which allowstheoretically high gains in terms of bit error rate.But in real transmission systems there are lots ofeffects due to nonidealities which decrease systemperformance, e.g. I/Q imbalances, DC offsets as wellas sample and/or carrier frequency offsets. Properestimation and correction of carrier frequency offsetsis of great importance especially inOFDMsystems.For future extensions of the IEEE 802.11a standardwith transmit diversity concepts, like Alamouticoding, the behaviour of these schemes under sucheffects has to be investigated.Our paper focusses on the impact of the carrierfrequency offset, which destroys the orthogonalityof the Alamouti scheme and leads to intersymbolinterference between the two alamouti coded symbols.This is shown mathematically and the resultsare compared with measurements in the 2.4 GHzISM band.Because a simple estimation of channel coefficientsand carrier frequency offset based on the originalIEEE 802.11a preamble is not possible in case ofAlamouticoded signals, we present a new preamblestructure, which is only slightly modificatedcompared to the standard. For estimating carrierfrequency offsets, two methods are presented, oneof them based on this new preamble structure. Wewill show in our simulation results, that even whenusing the new preamble for CFO estimation, whichallows an estimation accuracy similar to the one inthe SISO system, the loss due to estimation errors inthe MISO system is higher than in the single transmitantenna system.The paper is organized as follows: Section II dealswith the mathematical formulation of carrier frequencyoffsets inAlamouticodedOFDMsystems.Furthermore, we verify the results with measurementsin the 2.4 GHz ISM band. In section III,our carrier frequency offset estimation methods arepresented. Section IV shows some simulation resultsfollowed by a conclusion of the paper in section V.