A diversity strategy is efficient to reduce the fluctuation of communication quality caused by fading. In order to further maintain the communication quality and improve the communication capacity, this paper proposes a two-dimensional diversity approach by serially-concatenating spectral precoding and power normalized-differential space time block coding (PN-DSTBC). Spectral precoding is able to take benefit from a frequency diversity effect without loss in spectral efficiency. In addition, PN-DSTBC is robust against serious phase noise in an extremely high frequency (EHF) band by exploiting a spatial diversity effect. However, there is a problem that a naive concatenation degrades the performance due to the imbalance of equivalent noise variances over transmit frequencies. Thus, we examine an equalized PN-DSTBC decoder as a modified approach to uniform equivalent noise variances over frequencies. The performance evaluation using computer simulations shows that the proposed modified approach yields the performance improvement at any modulation schemes and at any number of transmit frequencies. Furthermore, in the case of 64QAM and two transmit frequencies, the performance gain of the modified approach is 4dB larger than that of PN-DSTBC only at uncoded BER=10-4.

This paper presents the performance of DSTBC when applied on the downlink transmission of WCDMA cellular systems in fast-varying time-dispersive channels. First, three DSTBC-WCDMA receiver architectures are proposed and they are: (1) the DSTBC Rake receiver for combined-code (D-Rake-C), (2) the DSTBC deterministic receiver for combined-code (D-Det-C), and (3) the DSTBC deterministic de-prefix receiver for combined-code (D-Det-DP-C). Detection can be divided into a correlator that combines descrambling and despreading, and a DSTBC decoder. The correlator is designed to perform signal separation of the multipath-multiuser signal via least-square (LS) estimation. To enable the correlator to perform signal separation at every block period, the long combined spreading and scrambling codes are divided into shorter codes. Then, the proposed receivers are theoretically analyzed in time-dispersive channels and multiple-user environment using the moment generating function (MGF) of fading distributions. For analyzing interference tolerance, the standard Gaussian approximation is employed. Finally, simulations are performed. Theoretical performance well matches simulated results. Among the three receivers, the D-Det-DP-C receiver has the best performance in time-dispersive channels with a maximum excess delay of 4 chips and a maximum Doppler frequency of 250 Hz. Results also show minimal performance degradation for fast fading channels with a maximum Doppler frequency of 1200 Hz. The best performance is obtained when the receiver has the information on the maximum excess delay and all users' spreading codes.

This paper proposes a maximum likelihood sequence estimation (MLSE) for the differential space-time block code (DSTBC) in cooperation with blind linear prediction (BLP) of fast frequency-flat fading channels. This method that linearly predicts the fading complex envelope derives its linear prediction coefficients by the method of Lagrange multipliers, and does not require data of decision-feedback or information on the channel parameters such as the maximum Doppler frequency in contrast to conventional ones. Computer simulations under fast fading conditions demonstrate that the proposed method with an appropriate degree of polynomial approximation is superior in BER performance to the conventional method that estimates the coefficients by the RLS algorithm using a training sequence.

This paper proposes low-complexity blind detection for orthogonal frequency division multiplexing (OFDM) systems with the differential space-time block code (DSTBC) under time-varying frequency-selective Rayleigh fading. The detector employs the maximum likelihood sequence estimation (MLSE) in cooperation with the blind linear prediction (BLP), of which prediction coefficients are determined by the method of Lagrange multipliers. Interpolation of channel frequency responses is also applied to the detector in order to reduce the complexity. A complexity analysis and computer simulations demonstrate that the proposed detector can reduce the complexity to about a half, and that the complexity reduction causes only a loss of 1 dB in average Eb/N0 at BER of 10-3 when the prediction order and the degree of polynomial approximation are 2 and 1, respectively.