Comparison of Data Detection Schemes in Space Division Multiple Access Systems

Two data detection schemes for the uplink of a space division multiple access (SDMA) system are compared in this paper. The first scheme separates the dominant wavefronts using a two-dimensional (2-D) spatial filter based on 2-D Unitary ESPRIT. The separated wavefronts are assigned to the users by correlating them with known training sequences. To obtain temporal equalization, all wavefronts assigned to a particular user are passed on to different diversity branches of a (single-user) Viterbi equalizer. The performance of this two step spatio-temporal filter (STF) is compared to a computationally more expensive scheme in which a joint multi-user maximum likelihood equalizer is applied to the impulse responses of all co-channel users. The impulse responses are estimated by exploiting the training sequences. Monte Carlo simulations for synthetic and realistic scenarios allow a performance evaluation of both schemes based on the uplink bit error ratio (BER). Different ways of adapting equalization and coding to the STF are discussed briefly.