An estimation algorithm with learning feature for an adaptive bit synchronizer.

Abstract

Description of an on-line optimal (in the sense of linear minimum mean square error) sequential bit synchronizer assuming minimal a priori knowledge of the incoming data. The adaptive synchronization algorithms, implemented on a small computer, are capable of accurately estimating pulse transition time, pulse amplitude, and pulse width by invoking the powerful Gauss-Markov theorem. Numerous simulations indicate that the performance of the synchronizer is satisfactory from the point of view of the amount of data required and the speed of convergence.