Coding Performance with Imperfect Erasures

Convolutional coding with interleaving is frequently employed to obtain low decoded bit error rates in the presence of intermittent interference. It is well-known that the performance of the code is enhanced if the jammed bits are identified and erased. However, this enhancement is substantially reduced if the process of identification is imperfect. In this paper we calculate performance in the presence of both incomplete erasures (jammed bits not erased) and excess erasures (unjammed bits erased). Results are obtained for codes of constraint length seven and rates ¿, ¿, and ¿.