LDPC Coded Adaptive Transmission over Fading Channels

This paper first develops a simple closed form expression for the bit error rate (BER) of low density parity check coded (LDPC) M-ary quadrature amplitude modulation (MQAM) signals as a function of the signal constellation, received signal to noise ratio (SNR), and LDPC code parameters. Based on this new expression, joint optimization of the SNR target and transmission rate is studied for LDPC coded MQAM signals to maximize the average spectral efficiency (ASE) subject to an average transmit power constraint over Rayleigh flat fading channels. Total or truncated channel-inversion strategies are exploited for the inner-loop power control, since data rate and outer-loop SNR target are adapted either to coded BER (CBER) or to the SNR with instantaneous CBER constraint. It is shown that the use of BER-based adaptation instead of SNR-based adaptation is even more beneficial in coded systems than in un-coded systems in terms of ASE performance. Indeed some of our numerical examples show that adaptation based on CBER has up to 2-2.5 dB power saving over SNR-based adaptation on the ASE of coded systems.