Code-aided quasi-maximum SINR algorithms for blind linear multiuser detection

The problem of blind linear data-symbol estimation and data detection for the air- interface adopting the wide-band direct sequence code division multiple access multiuser multi- plexing technique with the promising short-code configurations has been addressed. The superior interference-suppression ability of the code-constrained minimum output energy multipath- component estimation is utilised to develop three code-aided quasi-maximum signal-to-interference and noise ratio (SINR) algorithms. These algorithms operate in such a way as to maximise approximate measures of the output SINR, each having variations, especially in the adaptive implementation, because of different criteria employed. The 'quasi-maximum SINR' nature is because of the approximations employed. Extensive simulations indicate that all of these algorithms significantly outperform the existing code-aided blind linear algorithms at con- siderably low computational complexity. Moreover their adaptive versions exhibit very high level of desirable trade-off between convergence speed and steady-state performance at further reduced computational complexity.