Linear Network Coding, Correlation Analysis And Its Application In The Route Of Infection Research

The network coding(NC) used in the mobile wireless sensor networks(M-WSNs) can improve the transmission performance and the robustness of the network. But it also bring in the great uncertainty mistakes because the linear network coding strategies may cause the problems of the coding packages' linear correlation. The redundant coding must be existed because the M-WSNs can not be maintained the same topology for a long time.The traditional "n*n" matrix encoding can not adapt to the redundant coding model. So the extended-matrix coding is proposed. The traditional "n*n" matrix can not be extended directly because it can not ensure the linear independence of the coding packages. The ring routing(RR) network model and the epidemic routing(ER) network model has been established in order to solve this problem. Depending on this two models, the random linear network coding and the determined linear network coding of M-WSNs have been analysed. The result show that neither the random linear network coding nor the determined linear network coding can be suited to the real environment of the M-WSNs or they can not ensure the linear independence of the coding packages.Based on this conclusion, an analysis model has been build to solve the problem that can be bring from the common encoding policy and a priority vandermonde fast coding(PVFC) policy has been proposed. The PVFC can be adapted to the M-WSNs environment easily and efficiently. Finally, the discrete event simulation followed proved the correctness of the previous models and the excellent fast-delivery performance of the PVFC.