Blind Recognition of LDPC Codes Using Convolutional Neural Networks

Abstract

The Low Density Parity Check codes have been widely used in modern communication systems, such as 5G new radio and fiber optic communications. In order to balance the quality and rate of communication, both sides of the communication tend to use different codes depending on the channel conditions. Therefore, the blind recognition technology of encoders is receiving increasing attention. At present, the blind recognition methods for Low Density Parity Check codes has been extensively studied. However, most of these methods require accurate estimation of the channel and are therefore limited to specific application scenarios. In this paper, we propose a method for blind recognition of Low Density Parity Check codes using convolutional neural networks. This approach is more flexible than the existing methods and can therefore be quickly deployed to new systems. Simulation results show that a simple network can achieve better identification performance than the existing methods.