Application of fractionally-spaced decision-feedback equalizers to HF fading channels

The authors investigate various techniques for improving the performance of fractionally spaced decision-feedback equalizers (FSDFEs) when they are applied to HF fading channels. Because of the time-varying characteristics of these channels, a rapid tracking capability of the equalizer is essential. The authors focus on various techniques which reduce the misadjustment error of fractionally spaced equalizers (FSEs), and thus, equivalently, improve the tracking. An algorithm based on the modification of the recursive modified Gram-Schmidt algorithm is presented, which utilizes the low-rank property of the input autocorrection matrix to reduce the misadjustment error by about 2.5 dB when an optimum sampling phase is used. FSEs with smaller sampling rates also prove to be beneficial in improving the performance if the transmitted signal has a smaller excess bandwidth. The results are then extended to the FSDFE.<>