Analysis Of Estimation Of MIMO Channel And Performance Of Antenna Array Based On Compressed-sensing

Wireless channel estimation is an important part in the MIMO communication systems. As a novel theory of signal sampling, the frequency of compressed sening belows the conventional Nyquist sampling frequency. The theory of compressed sening makes full use of the inherent characteristics of MIMO wireless channel. To some extent, we can reduce the number of pilot and increase the spectral efficiency and further to improve the performance of channel estimators by this method. As an indispensable aspect in the field of signal processing of antenns array, direction of arrival can locate the target signal and provide the technical support of received signals. In recent years, the theory of compressed sensing can solve the DOA estimation in the perspective of sparse reconstruction. That is, we can achieve high-resolution DOA estimators only by a small amount of experiment data. This undoubtedly promotes the DOA technology of estimation towards a more mature direction. The main work of this paper detailed as follows.Firstly, we introduce the characteristics of propagation in MIMO wireless communication, which mainly includes the large-scale fading caused by the path loss and shadow fading, frequency-selective fading caused by the multipath propagation, time-selective fading caused by the time-varying channel, as well as spatially-selective fading caused by the spread of angular. Then, we elaborate the sparse representation of signal, the designation of measurement matrix and the reconstruction algorithm of signal. We also propose some application about the compressed sensing in the MIMO wireless channel and analyze the feasibility of this theory according to the features of multi-path channels.Secondly, we establish a mathematical model of the local clustering sparse multi-path channel and carry out a complete derivation from the mathematical point to validate the possibilities about the utilization of this model. Unlike previous pure intensive or pure sparse channel, local clustering sparse channel model can reconstruct the original signals more accurately while reduce the rate of error more effectively. Moreover, the novel smoothed LO-norm algorithm has been greatly improved in the aspect of mean square error, reconstruction accuracy, matching degree and complexity of computation. Besides, this algorithm is fast and not sensitive to the noise, also exhibits a good performance of robust. Results show that the local clustering sparse channel is practical and the smoothed LO-norm algorithm is efficient.Thirdly, we propose that the channel estimation and coefficients of equalizer can inhibit the negative impact which caused by the ISI signal sequence. An improved smoothed LO-norm algorithm is introduced, which avoids the discontinuities of cost function when the smoothed coefficient is relatively small in smoothed LO-norm algorithm so that the robustness of additive noise is strengthened and channel estimators are more accurate. What is more, we design a comb-shaped channel model which can improve the performance of ISI much better. In terms of channel equalization, tests show that the additional FIR prefilter can solve the problem of interference more efficiently. In the case of low complexity, the proposed ISLO-prefilter viterbi algorithm can reconstruct original signals more precisely, and also improve the speed of computing as well as the matching degree of signals. Meanwhile, we can acquire our signals without the loss of performance in the receiver to further analysis.Fourthly, we explain the importance of DOA estimation, and introduce the L2,0-norm algorithm with the MMV to solve the spatial angle estimation about narrowband signals and wideband signals. Simulation results show that the performace of L2,0-norm DOA estimation algorithm are very good in both cases and outperforms MUSIC algorithm and MNM algorithm. Meanwhile, previous studies are basically based on the ULA. However, although the ULA can identify the specific locations of signals, it may cause the phenomenon of phase ambiguity due to the relatively strong direction selectivity of ULA. As a result, we extend the study to UCA. Numerical results show that we can complete unambiguous identification of target positioning in the new antenna arrays. Furthermore, we analyse the performance of beamforming with the application of L2,0-norm algorithm. Simulation results indicate a precise enhancement of signals.