Diagnosing-Based Mixed /Sources Localization With Unknown Sensor Defect Arrays

摘要

In this paper, localization of mixed far-field and near-field sources in uniform linear arrays (ULAs) with unknown sensor defects is addressed. The existing methods regarding localization of mixed sources are generally based on the assumption of ideal array manifold, which does not take the sensor defects into account. However in practice, sensor arrays may have unknown sensor defects, i.e., some elements of arrays have errors of phase and gain, or are broken-down, and the performance of these algorithms may decay in this case. To this end, we propose a new method for localization of mixed far-field and near-field sources based on sensor diagnosis in ULAs with unknown sensor defects. The proposed method can be divided into two stages. First, we design a spatial sliding window for constructing subcovariance matrices, which are utilized to identify the defective sensors. Then, based on sensor diagnosis of the previous step, we can avoid the negative influence of defective sensors and construct a special covariance matrix in order to estimate the DOAs for both far-field and near-field sources. Based on the estimated DOAs, the corresponding ranges of near-field sources can be estimated by the (1-D) range function. By eliminating the disadvantage effect of sensor defects, the proposed algorithm can obtain good estimation performance with respect to the DOAs and the ranges of mixed sources. The effectiveness of the proposed method is demonstrated by some numerical simulation results.