Rapid optimization method for polarization code decoding parameters

Abstract

The invention provides a rapid optimization method for polarization code decoding parameters. The method comprises the following steps of firstly, starting from collection and sorting of sample data;then carrying out modeling according to the characteristics and the size of the sample data, and adopting a supervised learning and random gradient optimization method for training a network; then inputting the likelihood ratio obtained by calculation of the received signal into a radial basis function neural network model which completes the training and outputting M; and finally, initializing Lto M and executing a serial offset list decoding algorithm. According to the method, a radial basis function neural network technology and a polarization code decoding technology are combined, so thatunnecessary calculation operation is avoided, and therefore the coding complexity of polarization codes is greatly lowered.

Description

technical field

[0001] The invention belongs to the technical field of communication, in particular to a method for optimizing polar code decoding parameters of a serial cancellation list decoding algorithm by using a radial basis function neural network. Background technique

[0002] Polar code is a new type of channel coding proposed by E.Arikan in 2008. Polar codes are the first constructive coding schemes that can be proven mathematically to achieve channel capacity. When polar codes were proposed, Serial Cancellation (SC) decoding was also proposed. SC decoding can be viewed as a path search process on a binary tree. The SC decoding algorithm starts from the root node of the code tree and searches for the leaf node layer layer by layer. After each layer is expanded, the better one is selected from the two successors for expansion. There are two main characteristics of SC decoding. On the one hand, it has low complexity and simple decoding structure; on the other hand...

Images