DOA Estimation Of Almost-cyclostationary Signal

Direction-of-arrival(DOA)estimation is an important issue in array signal processing.Although conventional DOA estimation methods can achieve high resolution,they do not make full use of the information of the incoming signal,and has some limitations in enhancing the degrees of freedom,strong interference rejection and low SNR resolution.With the deepening of research,it is found that many processes encountered in practical applications are cyclostationary signals.The signals are divided into interest and interference signals by the almost-cyclostationarity(ACS).By exploiting the ACS,the performances which are not get by conventional DOA estimation algorithms,such as estimating the directiong of arrival selectively,filtering out strong interferences and obtaining more degrees of freedom,can be achieved.The existing cyclic DOA estimation algorithms mainly exploit the cyclostationarity of the incoming signals,while ignoring the more general property of signals--ACS and generalized almost-cyclostationarity(GACS).In this article,we further develop the ACS process and more general signal form GACS process.We fouces on exploit the ACS to improve the DOA estimation performance.The main work of the article include the following:Firstly,to address the problem of the conventional prior interferometer direction finding algorithm can only estimate the direction of a single signal,we proposed a new DOA estimation algorithm for multiple signals based on the ACS processes analysis.The proposed methods exploit the time-frequency structure of the ACS process via the cyclic correlation function.The performance analysis shows that our algorithms can bring many advanteges which are traditional interferometer direction finding algorithms can not reach,such as the noise robustness,the selective derection finding,multi-signal processing capability and higher DOA estimation accuracy.Secondly,by combining the co-array and the property of ACS,a DOA estimation algorithm based on co-array is proposed.Due to almost-cyclic correlation matrix,the method retains the advanteges of cyclic DOA algorithms.It not only reduce the sensitivity to interferences and noises,but also allow us to obtain more degrees of freedom than the physical sensors.As a result,this method can enhance both the degrees of freedom and the accuracy of DOA estimation.Finally,we study the GACS process,and apply the GACS signal model in the scene of relative motion between the transmitter and the receiver.Conventional methods assume the relative radial speed is constant and model the Doppler effects as a frequency shift of carrier.However,this assumption usually invalid due to the high-maneuverability of target and would failed if relative radial acceleration is large.Compared to the conventional algorithms,the proposed GACS model take full account of acceleration.Numerical simulation results verify that the new algorithm outperforms the existing method which modeled the Doppler effects as a frequency shift of carrier,especially in accelerating scenario.