Research On Channel Estimation Algorithm Based On Compressed Sensing For LTE-A Uplink

In order to satisfy the higher demand of IMT-Advanced and future communication,3 GPP starts the standardization study of LTE-A,which is considered as the 4G technology standard. Wireless communication environment has a lot of uncertainty and randomness,among which,the characteristics of multipath fading and doppler effect of wireless channel make the received signal serious distortion. To offset the effects of fading channel, typically we use the channel equalization and correlation demodulation technology in the receiver.However, it needs to know channel state information firstly for channel equalization. From this view, we can see the performance of wireless communication system is related closely with channel estimation. In this paper the LTE-A uplink channel estimation algorithms have been researched From the following several aspects.Firstly, based on analyzing the SC-FDMA technology of LTE-A uplink, the conventional channel estimation algorithms based on pilots, including LS,LMMSE,DFT-LS and SVD-LMMSE algorithm and the data interpolation algorithms are studied. In order to reduce the energy leakage, this paper proposes a DFT-based channel estimation based on frequency noise variance estimation algorithm, in which we estimate the CIR through LS algorithm then use the results in frequency domain to estimate the noise average power as the threshold to be compared with the whole time domain channel energy. In order to improve the performance with a fixed threshold, this paper proposes a dynamic threshold DFT-based channel estimation algorithm,in which we first estimate the SNR and set a dynamic threshold based the estimated SNR to reduce the noise. The simulation results show that the two proposed algorithms both have better performance than LS and conventional DFT-LS algorithms.Secondly, after analyzing the characteristics of wireless channel, LTE-A system transmission model and the compressed sensing (CS) theory, the signal model meets the application criterion of CS and the time domain channel impulse response could be estimated based orthogonal matching pursuit (OMP) algorithm. Then this paper proposes a novel OMP-based channel estimation algorithm for unknown channel length and sparse degree. The simulation results show that the proposed algorithm has better performance than LS in BER and MSE, and it could save the pilot expense and improve the spectrum efficiency.Thirdly, LTE-A uplink SU-MIMO and MU-MIMO schemes are studied in this paper.Because of the MIMO channel joint sparse feature, the compressed sensing theory could be applied in LTE-A uplink MIMO channel estimation, and compressed sampling matching pursuit (CoSaMP) reconstruction algorithm is selected. The simulation results show that the joint channel estimation algorithm based on CS can effectively estimate the MIMO channel.