Serial JEVA for efficient decoding in impulsive noise channels

The detection of signals corrupted by non-Gaussian impulsive noise (IN) is an important topic because the IN poses a practical limit on the achievable system performance of many real-world communication systems. When accurate IN statistics are unavailable at the receiver, it is desirable to mark the IN-corrupted signals as erasures for subsequent decoding. We previously proposed the joint erasure marking and Viterbi algorithm (JEVA), which combines the conventionally separate processes of erasure marking and Viterbi decoding (Tao Li et al., Proc. IEEE Workshop on Sig. Processing Advances in Wireless Commun., 2003). We now introduce an efficient serialized version of JEVA, called the serial JEVA (SJEVA), which achieves a much lower average complexity at the price of only a fraction of a dB loss in SNR relative to the original algorithm.