基于聚类优选自适应KNN的改进定位算法

Improved localization algorithm based on clustering optimization and adaptive KNN

针对室内复杂环境下,WiFi定位算法选取固定K近邻(KNN)会导致定位精度变差的问题,提出基于MeanShift聚类选取自适应KNN的混合相似度加权KNN(MWKNN)定位算法,并基于几何位置对自适应KNN进行动态优选。通过MeanShift聚类和几何位置动态优选自适应KNN进行加权KNN(WKNN)算法定位估计,削弱了含有较大误差的近邻点参与定位的影响,显著提高了算法的定位精度。实验结果表明:在3 m网格及3 dBm噪声标准差条件下,改进MWKNN定位算法的均方根误差为0.92 m,平均定位误差小于0.74 m;2 m精度下的概率达到96%。定位精度明显优于传统KNN和WKNN算法,同时提升了定位结果的稳定性。

Aiming at the problem that the fixed K-nearest neighbor(KNN) selected by the WiFi positioning algorithm in the complex indoor environment will cause the positioning precision to deteriorate, the mixed similarity weighted KNN(MWKNN) positioning algorithm based on the MeanShift clustering to select the adaptive KNN is proposed, and the adaptive KNN is dynamically selected based on the geometric position.The weighted KNN(WKNN)algorithm positioning estimation through MeanShift clustering and geometric location dynamic optimization adaptive KNN,it weakens the influence of neighboring points with large errors in positioning, and significantly improves the positioning precision of the algorithm.The experimental results show that under the conditions of 3 m grid and 3 dBm noise standard deviation, the root mean square error of the improved MWKNN positioning algorithm is 0.92 m, and the average positioning error is less than 0.74 m;the probability of 2 m precision reaches 96 %.The positioning precision is significantly better than the traditional KNN and WKNN algorithms, at the same time, improving the stability of the positioning results.