Research On Compressed Sensing Based Channel Estimation In OFDM Systems

Since wireless signal propagated in the channel will be affected by multipath fading, leading to the received signal distortion in amplitude, phase and frequency, so it is necessary to compensate the received signal using the channel information. This dissertation studies the channel estimation techniques based on orthogonal frequency division multiplexing (OFDM) systems.The traditional channel estimation techniques assume that the channel is dense multipath, and the channel impulse is sampled relying on the Nyquist sampling theorem. Due to the high sampling frequency required by Nyquist sampling theorem, we need to insert additional pilot, increasing the cost of the system and reducing the spectral efficiency. Through long-term research on the channel, it is found that the wireless channel often exhibit sparsity.Compressed sensing is a new sampling theory rising in recent years, and its sampling frequency breaks through the limitation of Nyquist sampling theorem. As long as the signal satisfy the condition of sparsity, compressed sensing theory can use a small amount of sampling data to recover the original signal accurately.This dissertation describes the basic characteristics of the wireless channel and common channel model. Then it analyzes the process of the OFDM systems and some commonly used techniques. On this basis, we review the traditional channel estimation techniques, and summarizes some advantages and disadvantages of the traditional channel estimation algorithm. Then detailing the compressed sensing theory,we pointing out the advantages of its use in channel estimation. After analyzing an example of OFDM channel estimation using Orthogonal Matching Pursuit(OMP) algorithm, the dissertation presents a new greedy algorithm: Orthogonal Multimatching Pursuit(OMMP) algorithm. In the dissertation, we do a quantitative analysis to the complexity of the OMMP algorithm, proving that it is more efficient than OMP algorithm and more suitable for the larger sparsity channel. Computer simulation confirms OMMP algorithm can reduce the running time under the condition of same accuracy estimation. Finally, the dissertation analyzes the channel model of amplify forwarding mode, do a channel estimation to it using compressed sensing, and pointed out that the strategy of the pilot distribution. Through computer simulation, the feasibility of the algorithm is verified.