A simple proof that time-invariant convolutional codes attaincapacity

摘要

It is well known that time-varying convolutional codes can achievethe capacity of a discrete memoryless channel. The time varyingassumption is needed in the proof to assure pairwise independencybetween the codewords. We provide a relatively simple proof that indeedtime-invariant convolutional codes can achieve the capacity without anyrestriction (albeit, the error exponent achieved by our proof may not bethe optimal)