Researches On The Rate Distortion Properties Of The Distributed Video Coding

LDPC codes are near Shannon limit forward error correcting codes. As a new trend of future communication, more and more wireless broadband transmission systems are adopting LDPC codes as their channel coding schemes because of the powerful error correcting capability and low decoding complexity of LDPC codes. Meanwhile,there are also new requirements for the LDPC codes following the practical implementation progress of LDPC codes in more and more wireless broadband transmission systems. The paper discusses the topics around the practical applications of LDPC codes in the wireless broadband transmission systems such as the code construction, research on the principles of decoding algorithms and lowering the implementation complexity of LDPC decoders.First, the paper makes a summary of the modern techniques of LDPC codes. This includes the state of art of LDPC codes, some fundamental knowledge of LDPC codes and some comparisons between the existing code constructing methods of LDPC codes. Besides, the paper summarizes the principles of decoding LDPC codes in the probability domain.Then, the paper studies the density evolution algorithm and improvements on the optimization of degree distribution. Using the density evolution algorithm, we could chase the changes on the probability density of messages during the iterative decoding process under some reasonable assumption. Obviously, it's a ground work for constructing good performance LDPC codes to make clear how to implement the density evolution algorithm as fast and precisely as possible. Subjecting to the problem, the paper proposes a LUT-based method to compute the density evolution of the message directly. By joint considering the symmetric property of the computational function, the LUT method could promote the computation efficiency greatly while keeping the high precision of computation. Upon the method, it's easy to get practical optimized degree distribution pairs by joint considering the construction of LDPC codes. This greatly supports the construction of good error correcting LDPC codes.Then, subjecting to the requirement of LDPC codes with short code length and low encoding complexity in some wireless broadband multimedia transmission systems, the paper studies the new methods of constructing LDPC codes. As a subclass of LDPC codes, eIRA codes have a serially encoding structure that is well suited for encoding, while keeping good error correcting performance. Based on this property, the paper proposes two different constructing algorithms, which are called grouped block eIRA construction algorithm and grouped pseudo-random eIRA construction algorithm. By simulation, we find the eIRA codes constructed from the two algorithms not only have a comparably error correcting performance, but also have comparably low encoding complexity.Finally, the paper studies the design of multi-rate decoder of QC-LDPC codes. By designing base-matrix based decoder architecture and a new message updating module of variable nodes, the decoder is easy to multiplex the decoding units when decoding different codes rates of QC-LDPC codes. Meanwhile, the decoder architecture is configurable to different QC-LDPC codes by utilizing the regularity of QC-LDPC codes. Since the scalable transmission is one of the trends of the wireless broadband transmission systems, the base-matrix based multi-rate QC-LDPC decoder has a prospectively practical value.