针对声矢量阵幅相误差及阵元位置误差校正问题,基于特征分解法,提出一种简单实用的有源校正算法。该方法需要合作信源的至少3个方位信息,根据声矢量阵的通道特征,利用特征分解法构造矩阵方程组,通过矩阵运算得到声矢量阵阵元位置和幅相误差参数,从而实现对声矢量阵的校正。大量计算机仿真表明该校正算法具有良好的声矢量阵阵列误差参数估计性能。
Consider calibration for location, gain and phase errors of acoustic vector sensor array(AVSA), a simple and practical active calibration algorithm based on eigen decomposition was proposed. The proposed algorithm requires at least three DOAs(Direction of arrival) information from the cooperation source. The channel feature of AVSA and eigen decomposition are exploited to construct matrix equations, the location, gain and phase factors of AVSA are obtained by matrix operations and AVSA can be calibrated by the factors. Extensive computer simulations show that the proposed algorithm has good estimation performance for array errors of AVSA.