Date Added: May 2012
The LDPC code, whose check matrix H is sparse, is widely utilized in the current communication systems due to its good performance that is near the Shannon limit. The girth length, which is the minimum circle length of the Tanner graph corresponding to the code, has a large influence on the decoding performance of a LDPC code. In the traditional algorithms, they try to enlarge the girth length of the whole code directly to improve the performance of the constructed code. These algorithms are referred as global circle removal algorithms. In this paper, a new algorithm is proposed, which first partition the LDPC code into the low node degree partition and high node degree partition.