The Study Of Channel Estimation In Doubly Selective Channel Based On Compressed Sensing

The accuracy of the channel estimation has great significance to the reliability of the communication system.In some scenarios like underwater communication,low-altitude unmanned aerial vehicle communication,vehicle-to-vehicle networking,the time-varying characteristic of the wireless channel was caused by the Doppler effect of electromagnetic waves,with the frequency selectivity of the multi-path channel,which constitutes the doubly selective characteristic of the wireless channel.It is difficult to estimate the channel accurately in doubly selective wireless channel.Traditional channel estimation methods,such as LS estimation,can not meet the performance requirement in the doubly selective channel.However,according to the linear time-varying theory,the doubly-selective channel can be expressed as a sparse signal in the delay-Doppler domain.For sparse signals,the compressed sensing theory can accurately recover the original signal by a small number of sampling points.In this paper,the channel estimation technique based on compressed sensing in doubly selective channel was studied.The simulation results show that the channel estimation method based on compressed sensing achieves a better performance than the LS estimation under the same pilot cost.On this basis,some improved schemes based on compressed sensing was proposed to improve the channel estimation performance and reduce the complexity of the algorithm.The main contents of this paper are summarized as follows:Firstly,the importance of the study on channel estimation is discussed.And the research progress about doubly selective channel theory,channel estimation method,compressed sensing theory and channel estimation based on compression sensing are summarized.Secondly,physical characteristics of the doubly selective channel are discussed,and the linear time-varying system is used to model the doubly selective channel.The influence of the doubly selective characteristic to channel estimation based on pilots is analyzed in OFDM system.The basic theory of compressed sensing,including sparse representation,sparse reconstruction are studied.And then how to use compressed sensing theory for channel estimation is discussed.On the basis of this,three methods to improve the channel estimation based on compressed sensing under the doubly selective channel are proposed.First,delay filed and doppler field oversampling are proposed to improve the channel estimation performance.Second,a new pilot pattern design method is proposed to improve the performance.Third,combining the compressed sensing with the BEM,the applicable scenario of different spreading base is discussed.Finally,the complexity of the compressed sensing channel estimation is analyzed,and two methods to reduce the complexity of the compressed sensing channel estimation are proposed in combination with the prior information of the channel.