基于FLANN和最小二乘的磁梯度计误差校正

Error correction of magnetic field component gradiometer based on FLANN and least-squares

在基于偶极子磁场分量梯度的水下磁异常定位方法中,三轴磁力计自身误差及两磁场坐标系配准误差等是限制水下定位精度的主要因素,因此有必要对其进行校正,补偿磁场分量梯度计测量值。建立了磁场分量梯度计的测量误差模型,提出了基于函数链接型神经网络(functional link artificial neural network,FLANN)和最小二乘法的磁场分量梯度计误差校正方法,给出了误差参数辨识及校正算法,数值仿真和实测数据证明了校正算法具有良好的收敛性,能显著地抑制磁场分量梯度测量误差,该校正方法为提高磁场分量梯度计性能提供了一种可行途径。

In the underwater magnetic anomaly localization method based on magnetic dipole field component gradient,the error of tri-axial magnetometer and registration error of two magnetic field coordinate systems are the main factors limiting underwater localization accuracy,so it is necessary to correct the errors and compensate the measurement data of imperfect magnetic component gradiometer.The error model of magnetic component gradiometer is deduced and established in detail.An error parameter identification and correction algorithm based on functional link artificial neural network（FLANN） and least-squares method is given.Numerical simulations and experiment data prove the good convergence of the algorithm,which can remarkably reduce the measurement error of magnetic field component gradient.The proposed correction method provides a feasible way to improve the performance of magnetic field component gradiometer.