A Study On Fountain And Convolution Coding In Deep Space Communication

According to the fully utilized in multi-media communication and deep space communication of digital fountain technology, the characteristic of rate-less and non-feedback information has gain apparently advantage. This paper has been focus on the digital fountain codes, improve the design of fountain encoding and decoding algorithm with different environments, it has certainly both contribution in theory and realistic. The research described as follows:First, the background of the project has been introduced. With the development of digital fountain codes, the process of most important fountain codes such as LT code and Raptor is also been analyzed.Second, the encoding and decoding process of fountain code has been analyzed, the performance and complexity of coding has been decided by the constructor of generator matrix in some way, and the degree distribution and self constructor have mainly influence on generator matrix. So different degree distribution of fountain codes had been analyzed in chapter 3, the factor which contributes to the constructor of generator matrix has also been considered. The information node chooses by random or average can also show difference performance. Simultaneously, PEG algorithm has been proposed to eliminate the short loop in generator matrix. Consider using the above algorithm to contribute better generator matrix which can effectively improve the decoding successfully probability and the reliability.Third, the researching focus on the decoding algorithm of fountain code and proposed the decoding algorithm for LDPC code to decode LT code. After analyzed the code situation after decoding, the BP decoding algorithm has been improved to decoding for those re-using codes. System code with soft-decoding algorithm has also been simulated.At last, the researching focus on the reality limits of fountains codes. The traditional LT codes can only using in erasure channels, the performance in un-erasure channel would drop rapidly. For applied the fountain codes in un-erasure channels, concatenated code has been proposed. The inner code was utilized to translated the un-erasure to erasure channel. LDPC and Turbo is considered the best coding closed to the Shannon limits, so the inner codes had been chosen as LDPC codes or Turbo codes. In the result of system simulations, although the concatenated code has not the same performance as LDPC and Turbo, but the reality limits of fountain no longer exist, and digital fountain codes can be used in more different channels.