OFDM Channel Estimation Based On Compressive Sensing

In mobile communication system, the application of orthogonal frequency division multiplexing (OFDM) technology is becoming more and more extensive because of its outstanding advantages. Considering the transmission characteristics of the wireless communication system and the wide bandwidth of the OFDM system, channel estimation technology is a very important step for OFDM system. For existing OFDM system channel estimation methods, when allocating pilot data it is generally required to meet the Nyquist criteria, which leads to the reduction of the flexibility of pilot distribution. In this case,Compressed Sensing (CS) is considered to be applied in the OFDM channel estimation, which can break through the limitations of the Nyquist sampling criteria. Based on the sparse characteristics of the wireless channel, a model is established to solve the problem of channel estimation in OFDM system by using compressed sensing technology and the effectiveness and accuracy of channel estimation can be improved by this method.In this paper, we first introduce the classification and the difference of the measurement matrix, then the Toeplitz measurement matrix is introduced, and an optimization is carried out to improve the performance of the measurement matrix: take the Toeplitz matrix of interval line, and then, in order to enhance the correlation of the column, the orthogonal transformation of the matrix is carried out. The experimental results show that the optimization of the Toeplitz measurement matrix has a decline in the recovery of the residual error, while the exact reconstruction probability has been improved.But the Toeplitz matrix has an unstable disadvantage because of its randomness,therefore, this paper considers the use of chaotic sequence to construct the measurement matrix, because the chaotic system has a good pseudo-random nature, and can improve the stability. This paper considers using the Logistic chaotic sequence to construct the measurement matrix, and the optimization of the method of singular value decomposition is proposed: after the Logistic measurement matrix singular value decomposition, the improved measurement matrix can be obtained by changing the characteristic value according to the mean value. The experimental results show that this kind of improvement can significantly improve the accuracy of the Logistic measurement matrix.Then in the signal reconstruction process, the OMP algorithm is improved, because the OMP algorithm only select one element has the maximum correlation with the residual when updating each index set, in this paper consider choosing two elements have the maximum correlation with the residual when updating each index set, the experimental results show that the improved algorithm has a great improvement on the reconstruction probability compared to the OMP algorithm.Finally, an OFDM channel estimation model based on compressive sensing is constructed,above mentioned observation matrices and the improved OMP algorithm are applied in the model,the simulation analysis shows that the channel estimation method based on compressed sensing technology is better than the traditional channel estimation method.Further, the optimized Toeplitz measurement matrix, the improved chaotic sequence and the improved OMP algorithm can further improve the performance of the channel estimation based on compressive sensing.