The Research Of Online Calibration Methods For Multiple Sub-Arrays

In practical applications, it is difficult to avoid some kinds of errors, such as sensors position errors, gain and phase errors and so on. Imprecise knowledge of array manifold due to these errors will seriously degrade the performance of the high-resolution direction-of-arrival (DOA) estimation methods. Therefore, it is necessary to calibrate these errors for providing accurate array manifold. Now, array calibration has become a hot research direction in a variety of fields rang from radar, sonar, radio astronomy, oceanography, geophysics to biomedical.In this thesis, we consider three cases:array composed of multiple subarrays, near-field calibration source with known directions and strong interference source. Based on these cases, we research methods of array calibration for sensors position errors and gain-phase errors.The main work is as follows:1. Establish the model of array errors. First, in order to do comparison, we create the model of ideal received signal. Then, we establish the model of sensors position errors and the model of gain-phase errors. At the same time, we analyze the performance of DOA estimation methods with array errors.2. Study two common methods of array calibration. According to the linear relationship between the array mainfold and the eigenvector corresponding the maximum eigenvalue, we introduce calibrating algorithm with calibration source. And we use the characteristic of subspace orthogonal to introduce the F-W algorithm.3. For array composed of multiple subarrays, we consider the situation of the near-field calibration source to research algorithms of array calibration. First, based on the situation of the near-field calibration source, we take advantage of the structure of multiple subarrays to obtain the model of array errors and propose an algorithm which can calibrate sensors position errors and gain-phase errors at the same time. Then, considering the defect that the algorithm is poor for large gain-phase errors, we propose an improved algorithm. Finally, through the matlab simulation, we contrast the performance of these two algorithms.4. We introduce an approach to estimate the array manifold in the situation of a strongly interference source. The method utilizes the noise subspace, the signal subspace of the covariance matrix for the received signal and the constraint that the first element of the array manifold is equal to one to achieve the purpose that the array mainfold can recover from the interference signal received by the array antenna. Then, we can provide accurate array mainfold for the practical applications, such as high-resolution DOA estimation.