Assigning parameter estimates for N-survivor processing

Abstract

In an attempt to improve N-survivor processing (NSP), which lies between per-survivor processing and tentative decisions, a new algorithm is proposed for assigning parameter estimates to survivors. The proposed algorithm considers the history of the survivors in a trellis diagram by evaluating the distance between the hypothesised input vectors that contain the path history. It will be shown through computer simulation that NSP employing the proposed algorithm can outperform the original version.