Performance of a sub-optimal metric for soft-decision Viterbi with bit interleaving in OFDM based systems

Abstract

In OFDM (orthogonal frequency division multiplexing) based communication systems, convolutional codes are used along with bit interleaving to counter the effect of burst errors. The receiver, on the other hand, performs de-interleaving and decodes the received bit sequence using the Viterbi algorithm (VA). One obvious choice for this type of decoding is to use the Hamming distance and hard decision VA. The performance of a metric based on the Euclidean distance, allowing a soft-decision Viterbi decoding on the received de-interleaved bit sequence, is studied. Unlike conventional soft-decision Viterbi decoding, demodulated bits are assigned soft metrics based on the corresponding received symbol. It is observed from simulations that bit error rate (BER) performance using this metric is very close to the conventional symbol based soft-decision Viterbi decoding which can be applied when bit interleaving is not incorporated. We also discuss a modified metric using the frequency domain equalizer (FDE) taps in the case of multipath channels, which yields better packet error rate (PER) performance.