Fixed complexity error pattern estimation

Recently, various techniques using cyclic redundancy check (CRC) codes for error correction have been proposed. In previous techniques, a small number of unreliable bits in a packet were toggled in order to change negative acknowledgement (NAK) into acknowledgement (ACK). The difficulty of using these techniques is that the worst case complexity is still high because the number of possible error patterns is 2^N_UR where N_ur denotes the number of the unreliable bits. In this paper, we define the optimality of a set of error patterns when the set size as well as the number of unreliable bits is fixed. We then show that the optimal set of error patterns can be found with feasible computational complexity even when N_ur is in the order of hundreds.