Research On Algorithm Of High Accuracy DOA Estimation And Tracking Based On Single Channel Array

Compared with multiple-channel array system,single-channel array system has some advantages such as small size,low cost and low system complexity.It has significant meaning for reducing direction-finding system size and its application in practice to study high-accuracy spatial spectrum technique based on single-channel receiver.Based on the frequency-domain model and time-domain model,this paper discusses the problems of DOA estimation and tracking.The main work and achievement of this paper are summarized as follows:1.Establish the frequency-domain and time-domain spatial spectrum estimation model based on single-channel receiver.Firstly,based on the sequential sampling structure,we consider the infinite data acquired by sequential sampling as a stochastic process and regard once sequential sampling as once realization of stochastic process,referred to as an example function of stochastic process.The frequency-domain model suited for far-field narrow-band signals is established by phase compensation.Then,using the features that the time-delay of signal between adjacent elements is unrelated to its bandwidth,the time-domain model based on time-delay compensation is built and the wide-band and narrow-band signal can be processed simultaneously in the model.Finally,simulation results verify the reliability.2.Aiming at the case of non-uniform Gaussian white noise,this paper proposes a power domain least square direction of arrival estimator.Firstly,the least square expression of error power is derived.Then,The simulated annealing genetic algorithm is used to search the optimum value of the objective function and get the estimation of DOA,which improves the search accuracy because that the algorithm have not only the good global control ability of genetic algorithm,but also the strong local search ability of simulated annealing.A closed expression of objective function is given by the estimator,which avoids the iterative calculation and error accumulation.The estimator can be applied to uniform Gaussian white noise and non-uniform Gaussian white noise simultaneously because that the expression does not contain parameters related to noise power.3.Based on the established time-domain model,high-resolution DOA estimation for wideband signal using single channel receiver is performed by means of Bayes maximum a posterior probability estimation method.Firstly,the maximum a posterior probability density function of state vector is derived.Then,the improved perfect sampling is utilized to search the optimum value of the function and finally the DOA of signal is obtained.Because the function contains the prior information of state vector and the parameters unrelated to estimation value are removed by integration instead of replacing by their estimation value,the algorithm can attain very high estimation precision.4.With regard to the case that the directions of signal vary slowly,this paper introduces the Cubature particle filter method proposed recently into DOA tracking with single channel receiver and achieves high-accuracy DOA tracking.This algorithm improves the construction of the importance density function of the basic particle filter and introduces the newest measurement information into the importance density function.Therefore,the distribution of particles extracted from the function is more close to the posterior probability density of the state and consequently improves the accuracy of signal DOA tracking.In the iterative process,to avoid the problem of numerical non-stability result from the fact that the symmetry and positivity of the error covariance matrix are destroyed,square root filter is combined with Cubature particle filter to form a square root Cubature particle filter algorithm.By performing the square root operation to error covariance matrix,its symmetry and positivity is preserved and its condition is reduced,so the numerical stability is improved.