Research On Joint Iterative Detection And Decoding Algorithm Based On Factor Graph

There exists multipath effect in the wireless communication systems, which causes channel distortion and intersymbol interference. This problem was solved well by the iterative detection and decoding technology derived from "turbo" iterative thought. The technology effectively combines the equalizer and decoder, and then dramatically improves the performance of the whole system. In "Turbo Equalization" technology, the maximum a posteriori probability (MAP) equalization algorithm is optimal, but it is hard to achieve because of its high computational complexity, so many researchers try to design an equalization algorithm which has a good tradeoff between performance and complexity.Factor Graph and Sum Product algorithm provides a new method to solve the problems in communication systems. In this paper, we used factor graph to deal with Turbo equalization problem, and then applied it into the iterative detection and decoding system based on LDPC codes. This frame not only gives a good equalization algorithm, also adopts Factor Graph integrate the equalizer and decoder into a graph, and explains the whole system in terms of a unified and joint view. The specific content is as follows:1. We described the definition of the factor graph and sum-product algorithm, derived the sum-product algorithm based graph model, and then used an example to depict it. Its applications and development prospects were introduced.2. The principle of turbo equalization was given, and the optimal MAP equalization algorithm and classical decoding algorithm were clearly discussed. Next, we researched using the factor graph to deal with the equalization problem in wireless communications, derived an iterative detection and decoding algorithm based factor graph in the turbo equalization system, and then simulated it on computer. The simulation results show the advantages of the equalization method on factor graph and the good tradeoff between performance and complexity over the sparse ISI channels.3. An iterative detection and LDPC decoding algorithm based on factor graph was proposed. The paper described the definition, the representation and the principle of coding of the LDPC codes, analyzed and compared its decoding algorithm and simplified decoding algorithm. After that, a joint iterative detection and LDPC decoding algorithm based on factor graph was proposed. The derivation and simulation results show not only this algorithm gets a good tradeoff between performance and complexity, but also factor graph with a unified view explains the whole system. Our study proved that the graph method has such nice characteristics as simple effectiveness and intuitive generality.