Research On Passive Localization Algorithms In Near Field And Mixed Fields With High Accuracy

Array signal processing is an important branch in the field of modern signal processing.It has a widely range of applications in both military and civilian fields.This thesis focuses on the research of passive source localization algorithms.In the source passive location model,the source model is divided into far field source location and near field source location by dividing the distance between the source and the sensor.When estimating the far-field source,DOA is the mainly estimated parameter.As for the nearfield source location,the estimated parameters include not only the angle parameter but also the distance parameter.In the near-field source localization and mix-field source location models,the proposed symmetrical nested array model provides new ideas for increasing the array aperture,improving the degree of freedom,and positioning accuracy.The main work and innovation of this thesis are as follows:(1)Improve the traditional near-field source localization method.By researching the classical near-field source algorithms,a near-field source localization algorithm based on sub-array partition is proposed.The uniform matrix is divided into two independent subarrays,and the angle and distance of two independent sub-arrays are estimated respectively,then get the actual parameter information.This algorithm avoids the parameter matching and 2-D search in the traditional near-field source localization algorithm,which improves the estimation performance.(2)A near-field source localization algorithm based on symmetrical nested arrays is proposed.A symmetrical two-level nested array is constructed based on the theory and method of two-level nested arrays.The symmetry is used to remove the distance information,and a matrix including only the angles is constructed.Then the vectorized de-redundancy method is used for the matrix to obtain a virtual uniform array with a larger aperture to perform angle and distance estimation.Due to the increased aperture,the estimation accuracy is higher,and the resolution is increased.(3)A new algorithm for separating mixed sources is proposed.By researching the correlation between steering vectors,an algorithm is proposed to separate far-field information from near-field information.The far-field information part is removed from the overall information,leaving only the part containing near-field information,to achieve parameter estimation.The algorithm can well separate the far-field source and the nearfield source,reducing the estimation error.(4)Expand the nested array theory to mix-field source localization algorithms.Firstly,the near-field algorithm based on nested arrays is directly applied to the mixed field.The far-field source and near-field source are classified according to the range of the near-field distance.On this basis,the method is further improved and another kind of method is proposed.This algorithm uses a cumulant difference method to separate the near-field source and the far-field source,and reduce estimation errors caused by angular ambiguity.Due to the increase of the array aperture and the degree of freedom,the parameter estimation error is reduced,and the number of sources can be estimated to increase.