基于混合优化算法的单分量磁性定位方法

A new magnetic localization method based on vessel′s vertical magnetic field

为解决现有磁定位方法的定位精度较大程度上依赖于定位参数初始值估计精度的问题,提出了基于混合优化算法的单分量磁定位方法。首先利用舰船垂向磁场的衰减特性构造了相关系数函数,该函数与测量面上磁场的拟合误差函数共同组成了一种新的多目标规划函数,然后将微分进化算法与方向加速法相结合形成了混合优化算法,再利用该算法通过多目标规划函数确定最优的定位参数解。实验证明,对于未安装定位系统的半合作舰船,该定位方法利用磁传感器所测目标的垂向磁场,能较准确地确定舰船相对于磁传感器的位置,定位结果稳定可靠,定位精度较高。

This paper presents a new mono-component magnetic localization method with hybrid optimization algorithm.The method constructs a new objective function which utilizes the vessel′s vertical magnetic field,and applies the hybrid of DE and Powell algorithm to find the optimum of the function.By means of ship′s model experiments,it is demonstrated that the new localization method can identify ship′s position relative to the sensors with high numerical precision and stability on the basis of that the vertical magnetic field of the quasi-cooperation ship with no tracking equipment fitted are known,and the new method can overcome the shortcomings of the exiting method by which the localization accuracy largely depends on the localization parameters′ initial estimate.