A Maximum Generating Function Method for Blind Recognition of LDPC Codes

Blind identification of channel codes has become a key area in the field of cognitive radio research, plays a crucial role in Adaptive Modulation and Coding (AMC). This paper introduces an algorithm for rapid identification of Low-Density Parity-Check (LDPC) codes from a set of candidates. The method combines the generating functions with posterior probabilities to evaluate the possibility of parity relationships. By calculating the maximum eigenvalue of the generating function, LDPC code parameters can be effectively identified. Then, the mean square likelihood ratio is utilized to amplify the eigenvalues. Finally, LDPC code parameter identification is completed by comparing the magnitudes of the eigenvalues of the parity check matrices in the candidate set. Simulation results show the effectiveness of this approach. Compared to existing algorithms, this method not only improves recognition accuracy but also enhances the recognition rate by 13%.