A New Method for LDPC Blind Recognition over a Candidate Set Using Kullback-Leibler Divergence

Blind recognition plays a very important role in both applications such as grant-free access by Internet of Things (IoT) devices and non-cooperative communication scenarios. Therefore, it has received more and more attention in recent years. In this letter, a new scheme is proposed to resolve the problem of blind low-density parity-check (LDPC) encoder recognition over a known candidate set. First, we measure the relationship between the received vectors and the rows of the parity-check matrices in the candidate set, which can be represented by different distributions depending on whether the parity-check relationships are satisfied or not. Then, we classify the LDPC encoder over the candidate set by employing Kullback-Leibler (KL) divergence metric for measuring the distance between these two distributions. Simulations show that the proposed algorithm provides better recognition performance.