Research On Fountain Code Encoding And Decoding Algorithm With Deterministic Delay

With the quickly changes of society,and the massive application of the modern technology of communication,people have higher requirements for communication experience.Applications represented by services such as ultra-high-definition video on demand,AR images,and VR video experience have been widely used,and this new file transfer Compared with traditional services,there are higher requirements for transmission rate,transmission bandwidth,transmission delay and other related parameters.The fountain code has the characteristics of no rate,no feedback,low coding complexity,etc.,which can well improve the problems of packet loss,misalignment,and congestion in the existing network channels.However,due to the characteristics of random coding,the fountain code has a large randomness in the decoding process,and the decoding delay jitter is too large,which greatly limits the application of fountain codes.Therefore,this article focuses on the existing fountain code compilation Research on code algorithms to improve the decoding effect of fountain codes.On the basis of analyzing and verifying the influence of different factors on the decoding performance of fountain codes,this paper studies the number of packets,code length,coding overhead,packet loss rate and other factors through theoretical analysis and simulation experiments,and also studies common BP decoding algorithms.The performance of the GE decoding algorithm was analyzed and verified,which laid the foundation for the subsequent work on improving the fountain code.On the basis of theoretical analysis,this paper improves the existing GE algorithm of fountain codes by using the theory of the largest connected component of the graph.First,the criteria for selecting rows to be eliminated in decoding was re-selected.The first step of elimination can be completed with lower complexity.Then,in order to reduce the computational complexity of the remaining steps,a triangular matrix is used to multiply the processed matrix.,The unprocessed matrix becomes sparse,and finally the matrix is transformed into a diagonal matrix with full rank by matrix elimination.Finally,the average time consumption of the improved decoding algorithm and the performance of decoding jitter are analyzed and simulated.Experiments show that the new algorithm can effectively reduce the decoding time and has lower decoding delay jitter.At the end of this article,in order to realize the fountain code design scheme and verify the actual transmission effect of the decoding algorithm,on the basis of the theoretical design scheme,a set of file transmission system is designed to test the erasure ability and time delay of the fountain code.performance.By simulating the packet loss of the channel,the accuracy of file transmission and the transmission delay under different encoding overheads are tested.