摘要
针对室内地磁定位中因地磁属性随环境干扰导致定位精度不高的问题,首次将蚁群算法引入室内定位的地磁指纹匹配过程,并提出了6种思路。通过在某工作区域与KNN算法进行实验测试与对比,发现:当不同位置间的属性完全相异,同一位置不同时间点的属性基本相同时,KNN算法定位精度较高;当不同位置间的属性相似,同一位置不同时间点的属性受环境干扰时,提出的地磁指纹匹配思路既可保持易识别位置的高匹配率,又可提高受干扰位置的识别率。
This study aims at coping with low positioning accuracy in indoor geomagnetic positioning due to the environmental interference with geomagnetic properties.Ant colony algorithm is introduced into the process of geomagnetic fingerprint matching in indoor positioning for the first time,and six ideas are put forward.By comparing with results from k-nearest neighbor(KNN)algorithm in experimental tests at a certain working area,it is found that the positioning accuracy of KNN algorithm is higher if the attributes are completely different between locations and the properties of the same location at different time are basically consistent,whereas using the proposed algorithms maintain the high matching rate of easily identifiable locations and improve the recognition rate of positions subjected to environmental interference if the properties of different locations are similar and the properties of the same location are interfered by environment at different time.
作者
郭燕莎
GUO Yan-sha(School of Information Technology Engineering,Tianjin University of Technology and Education,Tianjin 300222,China)
出处
《天津职业技术师范大学学报》
2020年第1期1-6,I0002,共7页
Journal of Tianjin University of Technology and Education
基金
天津市教委科研计划项目(20140810).
关键词
室内地磁定位
蚁群算法
信息素
最近邻算法
indoor geomagnetic positioning
ant colony algorithm
pheromone
k-nearest neighbor(KNN)algorithm
