基于极大似然估计法的磁力计误差补偿算法

Compensation algorithm of magnetometer based on maximum likelihood estimator

针对现有三轴磁力计误差补偿速度慢、需要外部辅助设备、磁力计和惯性传感器组合存在多传感器轴位敏感重合误差问题,提出了一种基于极大似然估计法(maximum likelihood estimation,MLE)的快速有效的磁力计误差补偿算法。根据传感器组合系统中误差来源建立测量误差模型,建立高斯分布的极大似然参数估计模型,用牛顿最优法解算出误差补偿参数,并给出求解理想初始值的算法。仿真数据显示,补偿后的磁力计航向角解算精度达到0.81°,相比补偿前精度提高94.9%。实验结果表明,该算法可简单快速的实现误差补偿,多传感器轴位敏感重合误差得到校准。

To solve the problems in existing Tri-axial magnetometer(such as,error compensation is slow,External auxiliary equipment is needed and the axial-sensitive coincidence error when combined magnetometer and inertial sensors),This paper put forward a fast-effective error compensation algorithm of magnetometer based on MLE.First,establishing the error model through analyzing the errors source of the system and then,setting up the parameter estimation model based on the Gaussian distribution; finally,we use the Newton optimization method to solve the compensation parameters as well as giving the algorithm to solve the initial values.Simulation shows that the heading measurement precision after compensation can reach 0.81°,and the accuracy is improved 94.9% compared to the previous.Experimental result shows that the errors can be compensated easily and fast,and the axial-sensitive coincidence error of the multi-sensor can also be calibrated.