On the bit error probability for constant Log-MAP decoding of convolutional codes with rate k/n

In this paper, the analytical results of the bit error probabilities of convolutional codes using constant Log-MAP decoding by closed-form are presented, and the analytical results are compared with the Max-Log-MAP decoding results. Furthermore, theoretical improvement employing the correction term, which corrects errors induced by the maximum approximation, is shown. The results show that the bit error probability of constant Log-MAP decoding is marginally better than that of Max-LogMAP decoding.