Non-binary coded modulation and iterative detection for high spectral efficiency in MIMO

Non-binary low density parity check codes (LDPC) have been shown to outperform their binary counterparts and attain the best-known performance among error-correction codes. At the same time, it has also been established that wireless communication using multiple-input multiple-output (MIMO) schemes dramatically increases system capacity and reliability when information symbols are appropriately coded and modulated across transmit antennas. Recent works have studied techniques for applying non-binary LDPC codes to MIMO systems and demonstrated near-capacity performance. However, the constellation sizes in the proposed approaches are limited by the Galois field (GF) size of the non-binary LDPC code, thus limiting the maximum spectral efficiency of the proposed joint modulation-coding approaches. In practice, the GF size may be limited by computational complexity reasons but high spectral efficiencies are desired. Thus, we study iterative coded modulation based techniques for applying non-binary LDPC codes to higher order modulations. These methods have been well-studied for binary codes but not for non-binary codes. It is demonstrated that performance of the technique is dependent on the mapping of GF symbols to constellations, but applying the appropriate mapping attains near-capacity performance. To reduce the error floor region for smaller field sizes, while maintaining steep waterfall curves, we combine a nonbinary-based accumulator with a parallel concatenated coding scheme.