Array Direction Of Arrival Estimation Based On Difference Operation On Phase-shift Of Source Signals

The array signal processing based on various sensor array is the modern signal processing technology which has developed rapidly in the last 30 or 40 years.As a representative and central problem in array processing,DOA estimation is the basis and key link to describe,locate,identify and track the target or unknown source and has been extensively applied to the great diversity of engineering applications including radar,sonar,wireless communication,radio astronomy and seismic prospecting,etc.DOA estimation is of great value in military and civilian field.Mathematical model of narrow band signal and common array structure in array processing is introduced firstly.On this basis,classical DOA estimation algorithms including MUSIC,root-MUSIC,ESPRIT,spatial smoothing MUSIC and Improved MUSIC(IMUSIC)have been studied and analyzed by numerical simulation.By summarizing strengths and weaknesses of existing DOA estimation methods in theory and performance,it is shown that existing methods are usually proposed based on unknown or uncontrolled signal characteristics of external signal and sensor noise,as a result,it is easy to appear that the signal hypothesis and the actual situation do not match;at the same time,in complex environment such as multi-dimension array and color noise background,signal modeling is difficult,the algorithm is clumsy and awkward which results in the poor method applicability.Due to SVD of the related data often involved,these indirect methods have suffered from computational burden,and furthermore,seem to be misleading and make the problem more complicated.Secondly,as for these problems,a new DOA estimation method based on difference operation on phase-thift of source signals which can be applied for independent source and coherent source has been proposed for uniform linear array(ULA).Based on the constant property of phase-differences of sources induced by array structure,selecting a sequence of received data correlation functions of which phase differences linear changing with corner mark of sensor position as research object,a general form of DOA estimation matrix equations has been constructed by successive difference operations on phases of sources.According to the formal characteristics of the coefficient matrix of matrix equations,a higher order equation for incident angle of source signals has been proposed.By the process of solving the matrix equations and extracting the signal direction parameter information,two methods named analytical method and numerical method have been summarized to respectively obtain analytical solutions and numerical solutions according to the number of source signals.Finally,simulations are performed to compare performance of DOA estimation equations set with different concrete forms,and the matrix equations set by which optimal estimation performance can be obtained has been selected.The simulations of estimation performance of proposed method itself and comparison with MUSIC and spatial forward backward smoothing MUSIC algorithm demonstrate that the proposed method can deal with dependent,mix and coherent signals well in low SNR situation.Especially for coherent source signals,the proposed method has better estimation performance than the traditional spatial forward backward smoothing MUSIC algorithm.