Capacity of Insertion and Deletion Channels

Theoretical estimates for lower bounds on the capacity of various insertion and deletion channels are calculated by considering the rates achievable using convolutional codes and sequential decoding. It is found that these limits fall below the rates that can be achieved by the use of watermark codes over similar channels. The sequential decoding procedure used is shown to be suboptimal, but it is expected that a decoder which was able to take account of the phase of received and transmitted sequences would lead to an improved lower bound estimate for the capacity. However, it is unknown how this would compare to the capacity estimation that watermark codes give.