Sampling Frequency Error Estimators for PRIME v1.4 systems

Abstract

The performance of PRIME v1.4 systems might be severely degraded by a sampling frequency offset (SFO). This work addresses the problem of the SFO estimation by means of the pilots transmitted in the header and payload symbols. Estimators based on the best linear unbiased estimation principle are proposed and the frame error rate (FER) achieved with them is assessed both in an additive white Gaussian noise (AWGN) channel and in a set of 174 links obtained from measurements. Results indicate that the estimator based on the header symbols suffices to achieve almost the same performance of the no SFO case. Gains are shown to be particularly important when only the highest PRIME channel is employed and uncoded DQPSK and D8PSK are used.