Study Of Decoding Algorithms Of Convolutional Codes

Convolutional code has shown its great advantages since it was proposed by P.Elias. Block convolutional codes and Turbo codes are two typical types on the basis of convolutional codes. In the future, the wireless communication system will need better performance to meet the development of communication business, which will lead more codes with better error performance to appear.The decoding algorithms of convolutional codes are mainly researched in this thesis. Firstly, the basic architecture of digital communication is introduced, and channel coding is mainly introduced as an important part of the basic architecture. Convolutional code is a classic type of code and has been used in error control fields widely. Secondly, the principle of convolutional codes and the detailed derivation of decoding algorithm of BCJR are stated, and the error performance are simulated and analyzed under different decoding algorithms such as Max-log-MAP, MAP. Thirdly, the Max-log-MAP decoder for convolutional code (1,15/13) is designed under the platform of the FPGA, in which the concept of Sliding-Window(including Pre-warming Window and Computing Window) is introduced and time division multiplexing of storage are adopted in order to shorten decoding delay and save hardware resources effectively.Finally, preliminary exploration of the application of SPA in convolutional codes is made in this thesis. Iterate SPA decoding model and non-iterate SPA decoding model are proposed. Furthermore, several convolutional codes with two or more shift registers are simulated respectively.