Compressive fault-tolerance network coding

Because many signals are correlated over temporal domain, this paper proposes a compressive fault tolerance network coding scheme for exploring the temporal correlation of signals. This scheme overcomes the drawbacks of network coding theory by injecting the compressed sensing theory into network coding, i.e., the scheme possesses well compression gain. Meanwhile, it can tolerate finite errors and reconstruct the original information with low distortion when the rank of error matrix (induced by errors) is lying in a reasonable bound. Finally, the performance analysis shows that the compressed sensing-based network coding scheme outperforms the conventional network coding scheme with significant improvement in robustness and performance.