The increased demand of low-latency and spectrally efficient wireless transmissions requires accurate performance prediction mechanisms. A widely used performance prediction method for coded multicarrier transmissions applying iteratively decoded error correction codes is based on the employment of the mean mutual information of the coded bit. A precise knowledge at the transmitter site of the log likelihood ratios distribution at the receiver facilitates the accurate performance prediction and unambiguous modeling of the decoder’s behavior in a multiphase transmission.
This paper proposes an analytical method to characterize the distribution of the received bits log likelihood ratios using a limited number of parameters. The paper also evaluates the accuracy provided by the proposed theoretical method by comparing its results to the ones obtained from measurements. The performance evaluation shows that the proposed method provides good accuracies both for transmissions over Gaussian and frequency-selective channels.