证明了随机网络编码的传输矩阵具有压缩感知观测矩阵的性质,提出了一种基于随机编码和压缩感知的网络编码方案,该方案利用了压缩感知求解欠定方程组的优点。在该方案中,随机网络编码的误码率可以依靠信息的稀疏性而降低。如果信息足够稀疏,那么网络的容量甚至可以超过限定了网络编码容量上界的最大流-最小割定理的理论值。给出了误码率的理论上界,仿真结果表明该方案可以达到较好的性能。
It is proved that the random network coding has RIP in compressed sensing.A joint network coding scheme based on random network coding and compressed sensing is proposed.The scheme takes advantage of the compressed sensing ability for solving the undetermined system of equations.In this scheme,the error probability of random network could be reduced due to the sparsity of information.If the information is sparse enough,the network capacity could be even beyond the theoretical value of Max-Flow Min-Cut bound,which is considered as the upper bound of capacity in network coding.The upper bound of error rate is derived and the simulation result indicates that this scheme could achieve fairly good theoretical performance.