Research On Joint Channel Estimation And Decoding Algorithm Based On Factor Graph

For high-speed mobile wireless communication which the channel stateinformation is usually unknown, especially for the multi-antenna communicationsystem, there often exists multipath effect and time-varying which cause channeldistortion and intersymbol interference under the time-frequency selective fadingchannels, and then affect the performance of the system seriously. Optimalperformance of receiver relies on exact channel information. So it’s necessary toestimate the channel accurately. Researches have shown that joint channel estimationand decoding technology is a good solution to the problem. Therefore it’s a hotresearch topic to obtain the optimal performance for the receiver system with acomplexity as low as possible.As effective tools for iterative signal processing, factor graph and sum-productalgorithm can achieve unity of algorithm on graph model with simple calculation.Based on the research, this paper proposes a joint channel estimation and decodingalgorithm using iterative message passing as the theoretical framework and factorgraph as analysis tool.First, this paper explains the principle of turbo iterative receiver and introductionof factor graph, analyzes the specific messaging rules of sum-product algorithm basedon factor graph.Furthermore, the different SISO systems under time-varying flat fading channeland time varying frequency selective fading channel have been described respectively.By factoring of the joint posterior distribution function, the unified factor graphdescription of the system is achieved. Combined with sum-product algorithm and theturbo iterative theory, this article deduces messaging rules between different nodes ofthe unified graph model. To solve the integral of continuous variable of channelparameters and to reduce the computational complexity, the paper deduces theparticle filter algorithm based on the bayesian theory. And then a joint channelestimation and decoding algorithm based on factor graph is proposed. The simulation results show the algorithm can obtain a good performance in a parallel iterative waywith lower complexity.Finally, based on the algorithm above, this paper gives out the unity factor graphdescription of iterative receiver under MIMO system with the factoring of the jointposterior distribution function under time-varying flat fading channel. The receiverconsists of two separate modules named channel estimation and LDPC decoding andthe information has been exchanged and updated with the parallel iterative processingway. And then this paper proposes a joint channel estimation and LDPC decodingalgorithm based on factor graph for MIMO system. Simulation results show that thealgorithm can approximate the optimal performance, and its parallel processing waycan not only reduce the time delay, but also be more convenient to hardwareimplementation.