The Capability of Error Correction for Burst-Noise Channels Using Error Estimating Code

摘要

In the recent years, error estimating code (EEC) was proposed to estimate the bit-error-rate (BER) of a packet efficiently with very low data redundancy, which has been shown to be beneficial for wireless communications. In this paper, we provide theoretical analysis on the capability of EEC, which shows that correcting corrupted data bits using EEC is feasible when the BER of a packet is low. We further prove in theory that burst errors are more detectable than random errors using EEC, based on which we propose an error correction scheme for burst-noise channels. The proposed scheme includes several procedures such as segmentation, error detection, assessment, and flipping operation, which runs in polynomial time and can correct burst errors with some probability. Our work extends the philosophy of EEC from "estimation without correction" to "estimation with probabilistic correction". Numerical analysis based on a real- world WiFi communication trace shows that the proposed EEC-based algorithm achieves over 40% recovery ratio for corrupted packets in practice.