Joint source-channel coding based on polar code

Compared with the traditional separated source-channel coding model, the performance of the joint source-channel coding will be better in the actual communication system. In addition, the typical decoding algorithms all belong to the algorithms of separated source-channel model. Therefore, the joint coding and decoding algorithm is particularly important and it will be recognized as of great significance in future communication. In this paper, in order to meet the future communication requirements, a scheme of polar code in joint coding is proposed. At the transmitter of the system, the source polarization is constructed by the Bhattacharyya parameter construction method and encoded by non- systematic polar code. For channel polarization, polarization weight construction method is used, which was independent of signal to noise ratio. A joint Belief Propagation (BP) decoder with a simple structure is adopted in the receiver, which includes a channel decoder and a source decoder. The joint decoder performs external iteration only through a few iterations, thus reducing the complexity and improving the efficiency of the communication system. Simulation results show that compared with the channel BP decoder, the performance gain of J-BP decoder is about 1.8dB.