On SNR Estimation Techniques for Turbo Decoding Over Uncorrelated Rayleigh Fading Channels With Unknown Fading Parameters

Three non-data-aided signal-to-noise ratio (SNR) estimator techniques are derived and compared for turbo decoding, in conjunction with block-coded binary-phase-shift-keying (BPSK) signals received over fully interleaved (i.e., uncorrelated) Rayleigh fading channels. These estimators, which do not require knowledge at the receiver of the channel SNR nor the power of fading amplitude, are optimized in terms of minimizing the estimation error by choosing an appropriate curve fit model and comparing their estimation errors, as well as their turbo decoding bit-error-rate (BER) performance. Turbo decoding performance evaluation results have demonstrated that the two SNR estimators, which use the ratios of the second moment to the square of the mean absolute value and the fourth moment to the fourth power of the mean absolute value of the received signals, respectively, result in comparable BER performance to the case of perfect SNR knowledge at the receiver. Furthermore, they outperform the third SNR estimator, which makes use of the ratio of the square of the second moment to the fourth moment, particularly in applications with relatively short code blocks.