Multiuser or joint detection has recently been receiving significant research interest because of its potential for the significant increase in system capacity and performance. Among the conventional multiuser detection schemes such as the minimum mean-squared error and successive interference cancellation-based ones, maximum-likelihood multiuser detection has the best system performance. However, the complexity of maximum-likelihood detection (MLD) increases exponentially in the number of users and constellation size. In this paper, a low-complexity MLD scheme based on the use of a sensitive-bits algorithm is proposed. It is demonstrated that the proposed method can greatly reduce the computational complexity with a minimal penalty in performance compared with the exhaustive optimal MLD scheme.

A novcl feature weight (FW) algorithm and an intcgratcd schemc cdled MMSE-FW-LA are proposed. The schcmc combines the minimum mean square error (MMSE)-based enhancemcnt, the FW algorithm and the log-add (LA) modcl compensation. Expcrimcntal cvaluations show that this schcmc can improve thc recognition performance significantly, cspecially at low SNRs.

A differential detection scheme for transmit diversity was proposed by Tarokh. which can achieve full diversity order without the requirement to estimate the channel state at the receiver. This paper investigates the potential of using multiple receive antennas for differential space time coded MPSK signals over correlated Nakagami fading channels. We also investigate the effect of the carrier frequency offset (CFO) and channel correlation on its performance and present some results on its maximal tolerable frequency offsets for different MPSK signals. The results have shown that the differential encoding transmit diversity is very robust to the CFO and channel correlation. Copyright

A new scheme for robust automatic speech recognition is presented. The scheme combines minimum mean square error (MMSE)-based enhancement with Log-Add (LA) model adaptation. Experimental evaluations show that the proposed MMSE-LA scheme can achieve significant improvement in recognition across a wide range of signal-to-noise ratios, especially in very noisy environments.