Research On Non-circular Signals' Parameter Estimation

DOA(direction of arrival)estimation based on spatial spectrum estimation algorithm have high resolution and high accuracy,which is therefore concerned and researched by the majority of scholars,which also has broad prospects in engineering application of electronic reconnaissance and electronic countermeasure,but the spatial spectrum estimation algorithm does not consider or take advantage of the properties of received signal.If the properties of the received signal are used in reasonable,this is equivalent to improve the efficiency of the use of information,and can improve the estimation performance of space spectrum estimation algorithm.Modulation signals with non-circular property are common in the fields of radar,communication and sonar,non-circular property refers to pseudo covariance matrix of the signal is not zero,the spatial spectrum estimation algorithm using the one-dimensional information can improve the resolution and precision of angle measurement.This paper mainly focuses on non-circular signals to expand a series of studies,and proposes non-circular real-valued tensor ESPRIT(Estimation of Signal Parameters via Rotational Invariance Techniques)algorithm and the modulation signals' non-circular rate estimation algorithm,the main contents are as follows:1.In order to improve the angle measurement accuracy of the spatial spectrum estimation direction finding algorithm,a two-dimensional DOA estimation method based on the uniform plane array is proposed.Firstly,the data matrix is extended to tensor space by studying transform relationships between tensor and matrix;secondly,using the property of the non-circular signals of maximum non-circular rate and the Euler formula,make received data tensor change into two real-valued data tensor of sine and cosine data tensor,and in accordance with the antenna array dimension to concatenate in each dimension respectively,the higher order singular value decomposition of real-valued data tensor after concatenated,and according to the low rank approximation to obtain signal subspace;Lastly,according to the selection matrix to construct rotation invariant relationship,to realize DOA estimation by using eigen-decomposition to rotation invariant relationship to estimate the phase difference of between each dimension sensor elements of the antenna array jointly.Through the computer simulation experiment,it is proved that the method can improve the direction finding performance,and the experiment results show that the tensor can improve the accuracy of the angle measurement.2.Aim to modulation signal with non-circular characteristic,a non-circular rate estimation method is proposed.Firstly,the non-circular rate of several common modulation signals is derived theoretically.Secondly,giving two kinds of non-circular rate estimation methods according to the sampling data,Under without considering the space phase shift between the sensor elements,the first method is taking advantage of the maximum likelihood way to calculate non-circular rate according to the definition of non-circular;Under considering the space phase shift between the sensor elements,the second method is obtaining the non-circular rate estimation method of constant non-circular rate signals by taking advantage of the Hilbert shift to reconstruct steering vector and by conjugate derivation to the function that according to the maximum likelihood estimation,on this basis to improved to obtain non-circular rate estimation method of non-circular rate with the modulation length variation signal.Finally,simulation experiments are made to verify the correctness of the estimation of the non-circular rate of the constant non-circular rate signal and the non-circular rate with the modulation length variation signal,and this way have good estimation performance.