Performance Evaluation of Decoupled Space–Time Delayed Decision–Feedback Sequence Estimation in Mobile–Radio Environments (*) J. C. M. MOTA, A. L. F. de ALMEIDA (1) , C. E. R. FERNANDES (2) , F. R. P. CAVALCANTI {mota, andre, estevao, rod}@gtel.ufc.br GTEL–UFC: Wireless Telecom Research Group, Federal University of Ceará, Fortaleza, Brazil. URL: hhtp://www.gtel.ufc.br Abstract–This work aims at investigating the performance of a decoupled space–time processing structure based on a delayed decision–feedback sequence estimator (D–ST–DDFSE), in the context of the enhanced data rates for GSM evolution (EDGE) system. The main idea in using decoupled space–time processing is to separate co–channel interference (CCI) reduction from inter–symbol interference (ISI) suppression. Consequently, all degrees of freedom of a space–time front–end are dedicated to treat CCI, leaving ISI to be suppressed by a temporal equalizer. Due to the 8–PSK modulation and the large delay spread values compared to the symbol period, optimum detection becomes too complex in the EDGE system, which makes DDFSE a promising scheme for ISI suppression. The performance of D–ST– DDFSE is analyzed through link–level simulations under the context of COST 259 channel models for Typical Urban (TU) and Bad Urban (BU) propagation scenarios. Improved performance of this D–ST technique over a conventional space–time equalizer is observed. I. INTRODUCTION The mobile–radio environment of incoming mobile communication systems are characterized by strong co– channel interference (CCI) and severe intersymbol interference (ISI), which are key factors that limit performance and capacity. Third generation systems as the enhanced data rates for GSM evolution (EDGE) are characterized by high data rates requiring the use of equalization at both ends of the link. Furthermore, tight reuse configurations lead to strong CCI that limits system capacity. In the uplink, the use of antenna array diversity at the base station constitutes a classical solution for suppressing CCI and combating multipath fading [1]. The ISI due to multipath may be suppressed spatially by the adaptive antenna array. However, the rich multipath of practical propagation environments may limit the performance and it would be required too many antennas to overcome the effects of ISI and CCI. By introducing some kind of temporal equalization generally improves performance. Therefore, space–time (ST) processing techniques explore both the spatial and temporal structure of the received signals to obtain full (path) diversity. The classical space– time processing structure is the space–time linear equalizer (ST–LE) where a temporal equalizer follows each antenna element in order to exploit spatial and temporal diversity [2] simultaneously. In this case the minimum mean square error (MMSE) criterion is employed for CCI reduction and ISI suppression. Enhanced ISI suppression is obtained by the use of a maximum likelihood sequence estimator (MLSE) equalizer following an adaptive antenna array. However, in rich multipath scenarios the problem of insufficient degrees of freedom degrades MLSE performance due to residual CCI at the output of the array. Furthermore, in ISI–dominated scenarios with small values of angular separation, mainbeam user paths may severely reduce output signal–to–interference–plus–noise– ratio (SINR) and degrade bit–error–rate (BER) performance. It is known that the optimum solution against ISI is the MLSE equalizer while an MMSE criterion is more robust against CCI. Thus, it is reasonable to state that it would be desirable to treat ISI with an MLSE equalizer, which is the optimum detector in the presence of ISI. Similarly, CCI is better combated with an MMSE equalizer. The idea of separating CCI and ISI suppression has been studied by several authors [3–7]. A decoupled space–time (D–ST) processing technique can make use of the individual advantages of an MMSE– based algorithm for CCI reduction and an MLSE–based algorithm for ISI suppression. This is done by separating CCI and ISI mitigation in two stages. A canceling filter is employed to generate a modified version of the training sequence that adapts the array, so that it cancels only CCI leaving all ISI structure to be suppressed within a temporal equalizer, whose parameters are obtained from the coefficients of the canceling filter. D–ST equalization was introduced in [8]. In this work we evaluate the performance of a D–ST processing structure based on a Delayed Decision– Feedback Sequence Estimator, namely D–ST–DDFSE, in the context of the Enhanced Data rates for GSM Evolution (EDGE) system. Due to the 8–PSK modulation and the (*) This work is supported by the Ericsson Research – Brazilian Branch under the ERBB/UFC.01 Technical Cooperation Contract. URL: http://www.ericsson.ufc.br (1) Under– graduate scholarship supported by CNPq – Brazil. (2) Graduate scholarship supported by CAPES – Brazil.