Wireless OFDM Channel Estimation Based On Compressed Sensing

Due to the sparse characteristics of wireless channel,channel estimation methods based on compressed sensing(CS)can obtain channel state information accurately with less pilot overhead.Therefore,CS based channel estimation is a research hotspot of wireless communication at home and abroad.When CS based channel estimation method is used,one can improve the performance of sparse channel estimation by pilot design.Based on the CS theory,this thesis studies the pilot design for OFDM and MIMO-OFDM systems.The main contributions are as follows:1)OgOMP algorithm is proposed to estimate OFDM sparse channels.The proposed OgOMP algorithm selects dictionary atoms based on minimizing residual errors,and refines selected atoms at each iteration.The simulation results have shown that,compared with gOMP,the sparse channel estimation of OgOMP algorithm has smaller channel estimation mean square error,channel estimation bit error rate and better robustness.2)The joint mutual coherence and total coherence piltot design algorithm is proposed for OFDM system.It is proved that the pilot which can achieve lower bound of total cohence is cyclic difference set,and the lower bound of total coherence is derived.The pilot design is modeled as a multi-objective optimization problem,in which the mutual coherence and total coherence are used in pilot pattern design and pilot power design respectively.The simulation results show that,compared with the pilot design methods which design the pilot only based on mutual coherence or total coherence,the proposed method has smaller channel estimation mean square error and bit error rate.3)The shift based pilot design method and the tradeoff pilot design method are proposed to design pilots of MIMO-OFDM system.The simulation results show that the shift based pilot design method has higher computation complexity but has the best channel estimation performance.4)A sequential joint criterion design method is proposed to design pilots for channels with common sparsity in MIMO-OFDM system.Based on the block CS,the pilot design problem of MIMO-OFDM system is modeled as a multi-objective optimization problem which minimizes interblock coherence,subblock coherence and total coherence.The pilot power allocation is modeled as a second-order cone programming problem.A sequential joint criterion design method is proposed to design the pilot pattern and pilot power.Simulation results show that the proposed method has the minimum channel estimation mean square error and bit error rate.Based on the compressed sensing theory,the theories and methods of pilot design are proposed for OFDM and MIMO-OFDM systems.The proposed theories and methods form a complete technical framework.