Research On Channel Estimation Of TFT-OFDM System Based On Compressed Sensing In Massive MIMO

Since large-scale multiple input and multiple output(MIMO)has better performance in related applications than other non large-scale systems,it is applied to many mobile systems,and orthogonal cross division multiplexing(OFDM)can also have better performance in some channels,such as frequency selective channels,such as making the spectrum efficiency more efficient.Therefore,the large-scale MIMO OFDM will become the key to improve the performance of wireless communication system in the future.The traditional channel estimation technique assumes that the channel is a dense multi-path dependent on the Nyquist sampling theorem to sample the channel impulse response.Because of the high sampling frequency required by the Nyquist sample theorem,we need more pilot,it increase the system overhead and reduce the spectrum utilization.Through the long-term study of the channel,people find that the wireless channel often shows sparsity.Appeared in recent years,the compressed sensing is different from the Nyquist sampling theorem,it points out that the high probability reconstruction of sparse signal only need a few sample point,its application gradually appeared in various fields.This paper studies sparse channel estimation based on compressive sensing theory.First of all,the compressive sensing signal reconstruction algorithm is studied,the orthogonal matching pursuit algorithm(OMP),the regularized orthogonal matching pursuit algorithm(ROMP)and the iterative hard threshold algorithm(IHT)are described in detail.The basic principles and algorithms of these algorithms are elaborated in detail.The basic flow of the algorithm is given,and the relationship between the signal reconstructed power of various algorithms and the number of measurements and the sparsity of the signal is given.The simulation analysis of various algorithms is performed.The results show that the performance of the OMP algorithm is the best and the performance of the IHT algorithm is the second best,the ROMP algorithm has the worst performance.Second,MIMO OFDM systems are studied under the time-frequency training channel estimation and channel estimation are introduced to use multi-channel wireless channel sparse common support,here is a sparse mutual support and channel sparse levels are obtained by using time domain training sequence,this paper introduces a frequency domain pilot model,and in the pilot mode using the sparsity adaptive-simultaneous orthogonal matching pursui algorithm(SA-SOMP)for channel estimation,the last of the improved algorithm simulation analysis,comparing with other modes of channel estimation.Finally,this paper mainly studied the channel estimation method of massive multi-input multioutput orthogonal frequency division multiplexing system,it is proposed based on compressed sensing of large-scale MIMO OFDM systems under the time-frequency joint channel estimation algorithm.For large-scale time-frequency training under the MIMO orthogonal frequency division multiplexing system model has carried on the detailed introduction,according to massive MIMO OFDM system under the channel has a sparse,puts forward an improved compression perception algorithm for frequency domain estimation-apriori information based block orthogonal matching pursuit algorith(AI-BOMP),the algorithm requires a combination of time domain channel estimation to get information.Simulation results show that the proposed improved algorithm has better performance than SOMP algorithm,traditional MMSE estimation method and LS estimation method.