HIPERLAN-like turbo coded multi-user detection assisted OFDM based interactive video telephony

Abstract

Abstract-Multiuser detection assisted, multiple transmit antenna based OFDM arrangements are studied in the context of HIPERLAN 2-like systems. It is demonstrated that the system's user capacity can be improved with the aid of unique spatial user signatures, hence supporting a multiplicity of users. The Maximum Likelihood Sequence Estimation (MLSE) detection algorithm outperformed the Minimum Mean Square Error (MMSE) scheme by about 5dB in terms of the required Signal-to-Noise Ration (SNR) and this performance gain manifested itself also in terms of the system's improved video performance. I. BACKGROUNDRecently intensive research efforts have been dedicated to combining Orthogonal Frequency Division Multiplexing (OFDM) [1] with multiple-antenna reception assisted cochannel interference suppression. Since unique user-specific transmit antennas associated with unique user-specific spatial signatures are used, these arrangements are also referred to as Space-Division Multiple Access (SDMA) schemes. One of the key issues is the separation of different users, which can be performed based on their spatial signature, based on the knowledge of the channel parameters. A multiplicity of algorithms has been proposed for performing the task of user separation. The so-called Sample-Matrix Inversion (SMI) [2, 3] algorithm's roots date back to the early years of development in this research area. Other related algorithms, which have recently drawn interest are for example the Successive Interference Cancellation (SIC) technique originally proposed in the context of Lucent's V-BLAST system advocated by Foschini et al. in [4,5]. These techniques were further considered by Vandenameele et al. [6] and by Münster et al. [7], respectively. Furthermore, a Parallel Interference Cancellation (PIC) algorithm has been considered in [7], while Maximum-Likelihood (ML) combining and potential complexity reduction strategies were discussed by Awater et al. in [13]. A similar contribution was presented by Speth et al. in [14], where in contrast to [13], the focus is on the problem of maximum likelihood soft-bit generation. A comparative study of these detection schemes was also presented in [7], which concluded that MMSE combining has the lowest-complexity and hence it is the least effective detection scheme. The best performance was exhibited by the ML combiner at the cost of a substantially increased complexity, particularly in the context of higher-order modulation schemes, such as 16QAM. The complexity of PIC and SIC was observed to be within these two boundaries, with an advantage in favour of SIC, specifically in high-complexity scenarios, such as for example in conjunction with four receiver antennas, potentially supporting four simultaneous users. Hence, we opted for focussing our further investi- gations on the MMSE antenna-array output combining-and the ML combining schemes, more precisely on the MLSE soft-bit generation approach proposed by Speth et al. [14], which was found the be more amenable to our turbo coded transmission ...