基于余弦相似度的加权K近邻室内定位算法

基于位置指纹的室内定位系统能够实现较高精度的定位,其中KNN(K-nearest neighbor)和余弦相似度定位算法原理简单且易于实现。但每种算法仅从单一限制条件进行匹配,导致定位精度不高。针对此问题,提出基于余弦相似度的加权KNN算法,并通过实验测试算法的定位性能。实测结果表明,该算法的定位精度高于传统定位算法。当AP数量为5时,平均误差为1.67 m。定位精度优于1 m的置信概率为42%,优于2 m的置信概率为88%,最大定位误差为4.3 m。

WEIGHTED K NEAREST NEIGHBOR INDOOR LOCATION ALGORITHM BASED ON COSINE SIMILARITY

Indoor positioning system based on position fingerprint can achieve positioning with higher precision. KNN and cosine similarity positioning algorithms are simple and easy to implement, but each algorithm only matches from a single constraint condition, resulting in low positioning accuracy. To solve this problem, this paper proposed a weighted KNN algorithm based on cosine similarity and tested the positioning performance of the algorithm through experiments. The results show that the positioning accuracy of this algorithm is higher than that of the traditional positioning algorithm. When the number of APs is 5, the average error is 1.67 m. The confidence probability of positioning accuracy better than 1m is 42%, and the confidence probability better than 2m is 88%. The maximum positioning error is 4.3 m.