Coded Modulation Schemes with Signal Shaping for High Spectral Efficiency Communication Systems

The main aim of this thesis is to investigate the effect of signal shaping on the performance of coded modulation systems such as iteratively decoded bit interleaved coded modulation (BICM-ID) and multilevel coding (MLC) systems, which are applying various signal constellations in association with different forward error correction (FEC) coding schemes like convolutional codes, turbo codes and low-density parity-check (LDPC) codes. In spite of that both BICM-ID and MLC schemes can approach the Shannon capacity limit of the constellation set, the channel capacity is not fully exploited due to the use of signal constellations with uniform distribution, and a so-called shaping gain is still achievable. The shaping gain can be obtained by employing signal shaping techniques. Two main classifications of approaches to achieve shaping gain are constellation shaping and warping. While constellation shaping produces signal constellations with Gaussian-like distribution, warping technique generates geometrically Gaussian-like signal constellations. Among the most popular constellation shaping techniques are trellis shaping, tree shaping and shaping by non-linear block codes. On the other hand, using warping techniques the points of rectangular and circular constellations are optimized so that more points are placed at low energy levels and less points at higher energy levels. In this work, different constellation shaping and warping techniques are combined with BICM-ID and MLC systems. It was shown that the optimized systems employing signal shaping can provide further improvements compared to the standard systems.