Research On Channel Estimation Techniques Based On Compressed Sensing In OFDM Systems

On the basis of study the sparsity of the wireless channel response in somedomains, this article redesigns the pilot-based channel estimation method on the use ofcompressed sensing. First, a brief and comprehensive introduction to the radio channelcharacteristics and the basic knowledge of the theory of compressed sensing is given.Then, in the context of OFDM systems, specifically introduced the virtual channelmodel, pilot structure and signal reconstruction algorithm. The simulation result showsthat compared to classical LS method, method based on compressed sensing have betterperformance even with less pilot number.Finally, with the use of priori information channel two strategies were proposed toimprove the performance of conventional reconstruction algorithms. First, we exploitthe approximate sequential sparsity of the channel in order to track it over a period ofseveral consecutive symbol blocks. This approach can yield an additional performancegain, but more importantly it can substantially reduce the computational complexity.The second strategy researches on channel estimation schemes by combining CS andKalman filtering. CS is introduced to gurantee the correctness of the estimation ofmultipath delay information, with that information Kalman filtering channel estimationin time-delay domain is studied. Simulation results are provided in the final to verifythat this strategy improve system performance greatly.