New Bounds on Convolutional Code Performance over the Bernoulli-Gaussian Channel

Abstract

This work deals with the performance of convolutional codes transmission over the Bernoulli-Gaussian impulsive noise channel with blanking, where symbols subject to a noise pulse are zeroed before being fed to the decoder. As the blanking operation is equivalent to erasing the corresponding bits, the computation of the exact erasure probability of recursive convolutional codes is first derived. Second, bounds on the error rate curves are proposed. Finally, Monte Carlo simulation curves of bit error probabilities are compared to the derived bounds.