A Research On Channel Estimation Based On Compressed Sensing For Broadband Wireless Communications

Accurate channel state information is essential to reliable information recovery. Channel estimation technique as one of the key issues in broadband wireless communications has been a hot topic for a long time. The traditional channel estimation methods based on Nyquist sample theory usually have heavy pilot overhead, low estimation accuracy and poor performance. Wireless communications is developing to have border bandwidth and higher speed. This will lead to a higher demand on the validity and accuracy of the channel estimate techniques.The breakthrough of compressed sensing (CS) theory in recent years brings more opportunities for the development of channel estimate. According to the inherent sparsity of the wireless channel, the CS based estimation methods can give lower pilot overhead, higher efficiency and better performance. Applying this technique to broadband wireless communications has important significance and prospects, and thus receives much attention.In this paper, we first introduce the background of the CS and broadband communications technologies, following the basic concepts, theories and principles involved. We focus on the orthogonal matching pursuit (OMP) algorithm in CS, and give a new one based on noise variance (NV-OMP) though analyzing the limitations of existing OMP algorithms. This new algorithm can detect the multipath number adaptively, thus apply to estimate non-integer multipath delay channel. NV-OMP algorithm gives stronger robustness and better performance. Then the sparsity in Doppler domain of time-selective channel is analyzed, as well as the influence of the pilot pattern on the channel estimation performance of the CS based method. This method is also applied to estimate the channel state information in MIMO systems. Simulation results show that the pilot number required for CS based method can increase nonlinearly with the transmit antennas. Finally, we make research on the delay-Doppler sparsity of the doubly selective channel over multicarrier systems. System simulation results show that the method based on CS outperforms the traditional least square (LS) algorithm with half pilots and more accurate estimation.