Array shape calibration using near field pilot sources with unknown distance

Abstract

In this paper, an array shape calibration method for a planar sensor array is proposed. A sensor array is used for array signal processing that is a technique for direction of arrival (DOA) estimation, the beamforming, etc. When a sensor array with errors in sensor position is used, performance of array signal processing deteriorates. Therefore, such errors have to be estimated and compensated. To estimate errors in position, two or more pilot sources with known DOAs are generally used. Especially, a calibration method using near field pilot sources can reduce effects of noise and reverberations. When near field pilot sources are used, precise distances from pilot sources to an array are required. In the case that distances including errors are used, performance of the position error estimation deteriorates. In this paper, we propose a method for sensor array shape calibration using near field pilot sources with incorrect distances. In the proposed method, errors in distances of pilot sources are firstly estimated by minimizing the mean square error (MSE) of sensor position. Then, sensor position is calibrated using the estimated distances. The proposed technique is achieved using two near field pilot sources. The performance of our method is confirmed by simulations of DOA estimation using a sensor array that is calibrated by estimated distances and position.