Skip to Main Content
In this paper, a novel blind decision-directed estimation approach is presented for systems using energy detection of pulse-amplitude-modulated (PAM) signals. The method provides an accurate blind solution for the estimates of the received signal energy and noise power spectral density which are required for the calculation of the symbol decision thresholds. Numerical results demonstrate that the proposed approach is able to outperform an existing nondata-aided solution. Furthermore, already for relatively short observation intervals, the error rate performance of the proposed method is comparable with the performance of a corresponding data-aided threshold estimator which, unlike the proposed blind approach, needs a redundant training signal.