A feasibility study of the realization of adaptive equalizers with microprocessors

In spread spectrum data transmission systems employing pseudo-random type signals and microprocessor realizations. FFT techniques are indicated for computational economy. This offers the possibility of combining the functions of channel equalization and character recognition. This study concentrated on the aspect of channel equalization as representative of most of the functions required in a fully fledged system. The results indicate that (I) even if only equalization is considered, the use of a bit slice processor supported by a fast hardware multiplier is imperative; (2) such a processor could alternatively provide a weighted DFT output instead of a filtered time sequence to that the subsequent processor may affect character recognition without reprocessing the filtered time sequence and (3) for the particular processor considered time domain filtering is about twice as fast as frequency domain filtering.