Maximum likelihood SNR estimator for coded MAPSK signals in slow fading channels

Abstract

In this paper, an accurate maximum likelihood (ML) signal-to-noise ratio (SNR) estimator is proposed for coded M-ary amplitude phase shift keying (APSK) signals over a slowly time-varying fading channel. The proposed estimator significantly improves the estimation precision at low SNRs by utilizing a posteriori probabilities of coded bits provided by channel decoder. Moreover, a methodology to calculate the Cramer-Rao bound (CRB) of the proposed code-aided (CA) ML SNR estimator for coded M-APSK signals is derived. Compared with moments-based estimators and the non-data-aided (NDA) Expectation Maximization (EM) based estimator for M-APSK signals, simulation results show that the proposed estimator exploiting a posteriori information out of low-density parity-check (LDPC) decoder or Turbo decoder performs better, especially at low SNRs. It is also validated that the performances of the proposed CA-ML SNR estimator for coded 16- and 32-APSK signals can closely achieve the derived CRB.