基于遗忘因子递推最小二乘的无人机在线磁补偿技术研究

Research of On-Line Magnetic Compensation Technology for UAV Based on Forgetting Factor and RLS

针对无人机飞行过程中,由于载体非平稳磁场干扰造成磁测航向角不准确的问题,提出一种结合遗忘滤波的递推最小二乘在线磁补偿方法。在分析了软、硬磁干扰机理和建立磁干扰模型的基础上,对常规递推最小二乘法加入遗忘因子修正,从而抑制磁场变化对磁补偿参数估计效果的影响。仿真和试验结果表明,在磁干扰较强且干扰磁场存在突变的情况下,此算法可以将磁航向角RMSE减小至3.822 0°。该方法计算量较小,能够有效适应机载条件下的变化磁场补偿,并达到较高的航向角精度要求。

In order to solve the problem of inaccurate magnetic heading angle caused by non-stationary magnetic interference during the flight of UAV,a recursive least square（RLS）method combining with forgetting filter is proposed for online magnetic compensation.Based on the analysis of the mechanism of the soft and hard magnetic interference and a magnetic interference model,the conventional RLS method is modified by forgetting factor,so that the influence of magnetic field variation on magnetic compensation parameter estimation can be suppressed.The simulation and experimental results show that the magnetic azimuth angle’s RMSE can be reduced to 3.8220°with the strong magnetic interference and the sudden change of the magnetic field.The method has less computation and can adapt to the variation of magnetic field compensation under airborne condition effectively,and achieves higher heading accuracy requirements.