Nonlinear coded QAM modulation for nonlinear channels and parallel turbo equalization for sparse channels

We study interference mitigation approaches for various channels. First, nonlinear binary codes with QAM modulation are designed to control symbol transitions over the symbol constellation. This concept is employed to develop spectral shaping codes for QAM modulations, and also works as a nonlinear intersymbol interference (ISI) reduction technique in a peak power constrained nonlinear channel. These nonlinear binary coded QAM\u27s are investigated with an outer DVB-S2 LDPC code in a data-predistorted satellite link under an in-band average output power constraint. The concatenated scheme outperforms the conventional DVB-S2 LDPC coded system in total degradation subject to the constraint. Second, for linear sparse interference channels, we investigate parallel trellis MAP turbo equalization which efficiently exploits the nonzero taps of the sparse multipath channel over a few parallel trellises. Maximum likelihood sequence detection (MLSD) based on parallel trellises assuming perfect residual ISI cancellation is shown to achieve the same asymptotic bit error rate (BER) performance as the MLSD in convolutional coded ISI-free channels (we also show this is the same performance as the MLSD in convolutional coded interleaved ISI channels). Furthermore, Monte Carlo simulation of the parallel MAP turbo equalization shows similar BER performance at high signal-to-noise ratio (SNR), providing a significant complexity reduction compared to single trellis