基于改进Fast-DTW的多维地磁序列匹配算法研究

Research on Multidimensional Geomagnetic Sequence Matching Algorithm Based on Improved Fast-DTW

传统单点的地磁匹配定位算法运算量大,且定位时效性及定位精度较差。针对该问题,采用基于序列的快速动态时间规整算法Fast-DTW实现多维的地磁序列匹配,并利用最长公共子序列LCS的有序聚类进行优化。离线阶段通过坐标转换法构建四维地磁指纹库,在线阶段进行实时地磁序列匹配,获得定位结果。实验结果表明,该算法的定位时效性及精度均有提升。

The traditional geomagnetic matching positioning algorithms based on single points have a large computational overhead and relatively poor positioning timeliness and accuracy.To address this issue,this paper uses the Fast Dynamic Time Warping(Fast-DTW)algorithm based on sequences to implement multi-dimensional geomagnetic sequence matching,and optimizes it using ordered clustering of the Longest Common Subsequence(LCS).During the offline phase,a four-dimensional geomagnetic fingerprint database is constructed by coordinate transformation methods.In the online phase,real-time geomagnetic sequence is matched to obtain positioning results.The experimental results show that the positioning timeliness and accuracy of the algorithm have been improved.