Reaserch On Density Distribution Scheme Based On Sparse Network Coding In Realtime Multimedia Application

In sparse network coding,a large number of zero coding coefficients are selected,that is,the coding is only executed in the subsets of the original packet,so as to generate a sparse decoding matrix at the receiving terminal and reduce the computational complexity.One of the important characteristics of sparse network coding is partial decoding,that is,decoding partial original packets before decoding the whole original packets by collecting partial coded packets.This technology can reduce the transmission delay and the average decoding delay of each packet in real-time multimedia applications.The set density value of tunable sparse network coding can be defined by tunable definition,and the density value can be increased dynamically in the transmission process,which can effectively improve the generation probability of innovative coding packet and reduce transmission delay.A sparse network coding linear density distribution scheme based on absorption Markov chain is proposed to solve the problem that the fixed density value in the sparse network coding cannot be adapted to each stage of transmission,which leads to low generation probability of the innovative coding packet.According to the characteristics that the number of original packets required for the last transmission stage in traditional sparse network coding is approximate to that of random linear network coding,combined with absorbing Markov chain model and matrix rank probability model,state transition probability formula is put forward.And the rank increment or non-zero column increment is organically combined with the absorption state in the absorption Markov chain model in the decoding matrix,the original packet recovery probability formula is proposed further.The density value is adjusted under the premise of the probability value maximization to obtain the density value suitable for the characteristics of this stage.Simulation results show that under the same conditions,the proposed scheme has lower transmission delay and higher system throughput,and achieves better transmission performance under the condition of lower transmission delay.The existing tunable sparse network coding scheme does not have a high degree of integration with the actual network scene,and the factors such as bandwidth limitation and change of packet loss rate of link are not taken into consideration comprehensively,which leads to the difficulty in selecting the optimal density.Proposed in this paper,aiming at the above problems based on adjustable sparse network coding in real-time multimedia network linear density distribution scheme(LDDS)and phase density distribution scheme(PDDS)based on adjustable sparse network coding,two scheme adopts adjustable sparse network coding,set different density distribution function,based on the current obtaining-information node receiving feedback decoding matrix rank.Combined with the change law of rank and density in the budget comparison formula,the optimal density suitable for the current stage was selected,and the probability of innovative coding packet generation was improved by increasing the density gradually to ensure the decoding efficiency of nodes.The simulation results show that under the same conditions,the average number of decoding layers of LDDS is slightly better than that of PDDS,and the transmission mode based on generation is more suitable for real-time multimedia network,which can effectively improve the transmission efficiency.