In order to defeat both intersymbol and multiple access interference arising in code division multiple access systems, numerous multiuser detection techniques have been investigated . Actually, the computational complexity of the optimum maximum-likelihood multiuser detector (MLMD) increases exponentially with the number of users, therefore it is infeasible. We propose two suboptimum versions of MLMD, especially designed for down-link connections, which are based on a computational efficient reduced-state technique. It has been found that these algorithms have a significant performance gain as compared with other suboptimum joint detection criteria, such as zero forcing and minimum mean squared error. Moreover, they have a computational complexity which is proportional to the number of users.

We propose two reduced state versions of maximum likelihood multiuser detector (MLMD) for down-link connections in time division-code division multiple access (TD-CDMA) systems. Simulation results have shown that these algorithms perform substantially better than other widely used suboptimum detection techniques. Moreover, they have a computational complexity proportional to the number of users.

In the hybrid multiple access technique time division-code division multiple access (TD-CDMA), both multiple-access interference and intersymbol interference (ISI) arise. In order to combat the overall interference, we propose four efficient joint detection schemes based on zero-forcing and minimum-mean square error criteria. By exploiting the Toeplitz-block structure of matrices and the asymptotic equivalence between finite-order Toeplitz matrices and circulant matrices, most computations can be carried out very efficiently through extensive use of discrete Fourier transform (DFT) and independent DFT (IDFT) transforms. Performance results based on the UMTS scenario are presented.