| 摘要: |
| 传统的磁干扰补偿方法在静态或已知的地磁环境下表现良好，但在环境磁场动态变化的情况下，固定补偿系数可能不再适用，导致补偿效果大幅下降。针对这一问题，提出了一种自适应磁干扰补偿方法。当载体进入新的磁测环境时，根据传感器实时采集的磁场数据，利用鞍点优化的思想对标定的补偿系数进行建模，并结合近端梯度原始对偶算法更新优化框架下的对偶变量以及原始变量，逐步逼近最优解，实时更新补偿系数，确保在不同环境下的磁干扰补偿精度和效率。为了验证所提磁干扰补偿方法的有效性，使用水下载体平台数据和麻省理工学院(MIT)的飞行数据进行分析。结果表明，在不同地磁干扰背景下，此算法能够有效应对多样化的外部磁场干扰，与传统固定补偿系数方法相比，补偿后的磁干扰标准差(STD)降低了57.2%以上，显著降低了载体磁干扰水平。 |
| 关键词: 自适应磁干扰补偿 鞍点优化 近端算子 原始对偶 |
| DOI： |
|
| 基金项目: |
|
| An adaptive elastic network magnetic interference compensation method based on saddle point optimization |
| YAN Qicheng,ZHAO Yinan,LIU Zhaoting,WU Han |
| (School of Communication Engineering, Hangzhou Dianzi University, Hangzhou 310000, China) |
| Abstract: |
| Traditional magnetic interference compensation exhibits satisfactory performance in static or known geomagnetic environments, but the fixed compensation coefficients may no longer be applicable in cases where the ambient magnetic field undergoes dynamic changes, leading to a significant decrease in the compensation effect. To overcome this problem, a novel adaptive magnetic interference compensation method is proposed. When carriers enter a new magnetic environment, saddle point optimization principles are used to model the compensation coefficients calibrated according to real-time magnetic field data from sensors. Meanwhile, the proximal gradient primal-dual algorithm is combined to iteratively update the dual and primal variables within the optimization framework, facilitating convergence to the optimal solution and continuous adjustment of the coefficients. This approach significantly improves the accuracy and efficiency of magnetic interference compensation in varying environments. To demonstrate the effectiveness of the proposed method, data from underwater platforms and aerial data provided by the Massachusetts Institute of Technology (MIT) are analyzed. The results show that the proposed algorithm significantly improves the resilience to various external magnetic disturbances, reducing the standard deviation (STD) of compensation by more than 57.2% compared to the traditional fixed-coefficient method, thus significantly minimizing the magnetic interference on the carrier. |
| Key words: Adaptive magnetic interference compensation Saddle point optimization Proximal operator Primal-dual |
