Detection and Decoding Algorithms of Multi-Antenna Diversity Techniques for Terrestrial DVB Systems

This PhD dissertation analyzes the behavior of multi-antenna diversity techniques in broadcasting scenarios of TDT (terrestrial digital television) systems and proposes a low-complexity detection and decoding design for their practical implementation. For that purpose, the transmission-reception chains of the European DVB-T (Digital Video Broadcasting – Terrestrial) and DVB-T2 standards have been implemented over which diversity and MIMO (multiple-input multiple-output) techniques have been assessed through Monte Carlo simulations. On one hand, the most important multi-antenna diversity techniques such as CDD (cyclic delay diversity Alamouti code-based SFBC (space-frequency block coding) and MRC (maximum ratio combining), have been evaluated in a DVB-T system over both fixed and mobile Rayleigh and Ricean channels. With the DVB-T2 standard release, multi-antenna processing has actually been introduced in digital television systems. The distributed SFBC configuration proposed in DVB-T2 is analyzed from a performance point of view considering different propagation conditions in an SFN (single frequency network). On the other hand, error-performance and detection complexity analyses of 2×2 FRFD (full-rate full-diversity) SFBCs are carried out for last-generation DTV (digital television) systems. The use of channel coding based on LDPC (low-density parity check) codes in new standards such as DVB-T2, involves a soft-output MAP (maximum a posteriori) detection which results in an increase of the detection complexity. In order to study the FRFD codes behavior in such a BICM (bit-interleaved coded modulation) scheme, the Golden code, which achieves the maximum coding gain, and the Sezginer-Sari code, which has a lower inherent detection complexity as an expense of sacrificing performance gain, have been chosen. Using LSD (list sphere decoder) detection, BER (bit error rate) performance and computational cost results are provided for TDT scenarios. In order to overcome the variable complexity of the LSD, LFSD (list fixed-complexity sphere decoder) detection is proposed for practical implementations. A redesign of the previously proposed LFSD algorithm for spatial multiplexing MIMO systems has been performed for FRFD SFBCs with close-to-LSD performance. Furthermore, an analysis of the number of candidates is carried out in order to maximize the efficiency of the algorithm. Due to its fixed complexity, the novel algorithm can be fully pipelined making feasible a realistic implementation in chip.