Search

Pair-wise error probability (PEP) analysis for (pre) coded optimal framing, orthogonal frequency-division multiplexing (OFDM) when optimal training is employed for CSI acquisition is considered. It is observed that the use of optimal linear minimum mean square error channel estimator does not reduce the diversity but causes loss of coding gain and, more importantly, code design criteria on the presence of channel estimation error (CEE) do not change.

A novel algorithm is proposed for finite impulse response filtering and signal processing applications. The algorithm is derived by examining and modifying the asymptotic performance of gradient root mean square estimators, gradient estimators with dead-zone thresholds and optimal bounding ellipsoid algorithms. It is capable of operating efficiently in nonideal environments.

Autoregressive parametric spectra are estimated by modifying block covariance algorithms (BMCA) which are efficiently implemented in a frequency domain. The energy weighted combined forward and backward estimate is obtained by the development of another block modelled algorithm. The newly developed algorithm is known as the energy weighted BMCA, EWBMCA. Its performance is similar to that of Nikias and Scott's weighted covariance method.