Blind Recognition Of Turbo Codes And Compressed Sensing System Based On LDPC Codes

LDPC codes and Turbo codes are typical example of modern error correcting code, and have high performance to approach to the Shannon Limits. A sparse parity check matrix of LDPC codes can uniquely represents a kind of LDPC codes, the sparsity of parity check matrix determines the error correction performance of LDPC. Compressed sensing system can simultaneously aquire and compress information by nesessary compress matrix. Tt can save storage and contain enough information needed. Construction of reasonable and effective compressing matrix plays a key role on acquisition and recovery of signals. This thesis researches on the compressed sensing system of hyperspectral remote sensing datas by using parity check matrix of LDPC codes as the compressing matrix.Turbo codes use the special structure of coding and iteration of soft information between multiple decoders to realize pseudo random structure. Turbo codes have better performance than other codes and have been applied into many communication standards. Thus Turbo codes has profound influence on contemporary coding theory and they are worthy of extensive attention and research. Turbo codes have been widely used in land and deep space communications, as well as in military communications systems because of its complex structure. Therefore, blind recognition of Turbo parameters in intelligence communication, no-cooperative communication and electronic warfare fields is particularly important.Main work of this thesis is as follows:(1) This thesis proposed a compressed sensing system of hyperspectral remote sensing image based on LDPC codes. The system uses a sparse LDPC code parity check matrix as hyperspectral remote sensing image compression matrix, and uses orthogonal matching pursuit algorithm for image restoration.Compared with using the Gaussian matrix, using LDPC compressed matrix obtain better restored image quality.(2)This thesis first researched Turbo coding structure in3GPP standards and CCSDS standards, and simulated these two coding structures by C language. These two coding structures mainly consist of two intertwined3GPP standard algorithms and8-state recursive systematic convolutional encoder implementations, CCSDS interleaving algorithm and16-state recursive systematic convolutional code encoder implementations of the standard. Then the maximum a posteriori probability decoding algorithm corresponding3GPP standard coding structure is realized.(3)This thesis studied blind identification algorithm of zeroed Turbo codes. It realized recognition of the3GPP standard Turbo code length, interleave length, starting point and the code rate which employs matrix analysis method based on gaussian elimination.