The DOA Estimation Based On Order-recursive Least Squares

Array signal processing is extensively concerned by scholars because of its advantages in signal detection and feature extraction.It is applied to a great variety of engineering fields,such as radar,sonar,wireless communication,geological exploration and so forth.Array signal processing is not only a research content with important national defense security needs,but also a variety of civil and people's livelihood values.The direction of arrival(DOA)estimation has been a representative and conventional topic in array processing.Most existing DOA estimation methods need to know the number of external incident sources as a precondition,while the actual number of sources is often unknown or changeable,which needs to be provided additionally.Hence the more general and effective estimation method by which the DOA can be still estimated without giving the number of sources has been highly recommended in practice.In addition,the number of sources and DOA estimation are also affected by other signal properties on the source,such as the uncorrelated / coherent properties of the source signal,which should also be considered.The research contents of this paper are as follows:This paper takes the one-dimensional uniform linear array(ULA)as the research object,and proposes the spatial difference operation to extract the source phase difference,and organically integrates the phase-shifting difference operation with the stationary property of the source signal and noise.After theoretical derivation,the linear equations and root-finding polynomials satisfying the source phase difference are obtained.According to the linear equations of source phase difference,the order-recursive least squares(ORLS)method is introduced to study the DOA estimation problem when the number of sources is unknown.The main contents of this paper are as follows:1.Based on the phase-shifting forward difference operation(FDO)and backward spatial difference operation(BDO),the linear equations of source phase difference are derived,and the DOA estimation method is proposed by using the combined equations of the two equations.The system of linear equations of the source signal phase difference is also called the modified Yule walker(MYW)system of equations.2.According to the linear equations which satisfied by the source phase difference,it is proposed that the source signal model of one-dimensional ULA is a degenerate spatial ARMA process.The combination equations are applied to study the detection ability of the maximum number of sources in one-dimensional ULA model,the applicability of the combination equations and DOA estimation method for uncorrelated and coherent sources.3.The ORLS method is introduced into DOA estimation.The least square error and least square estimate(LSE)of each iteration are calculated by the way of order recurrence.The least square error of iteration is calculated by order recursive.In this paper,the angle imaginary part of the source estimation is considered to express the recursive convergence characteristics.4.The numerical simulation and analysis of the proposed DOA estimation method are provided.The research combines the physical characteristics of array signal processing with the signal characteristics,and introduces the recursive least square method,which provides a new idea for DOA estimation under the condition of unknown number of sources.