Statistical Approach for Blind Recognition of Narrow-sense BCH Codes

In this paper, a new algorithm for blind reconstruction of generator polynomial for narrow-sense Bose-Chaudhuri-Hocquenghem (BCH) codes is proposed. The proposed algorithm uses the property that the number of valid codewords is differed with respect to the dimension of code. Furthermore, by following the channel cross-over probability of binary symmetric channel (BSC), the probability that a received codewords is the valid codewords of a BCH code is calculated and used in proposed algorithm. Through the hypothesis test which use the number of valid codewords, the BCH code used at the transmitter is distinguished at the receiver and then the generator polynomial can be reconstructed. Simulation results confirm that the proposed algorithm has much better recognition performance than the previous method.