参考点对WiFi位置指纹算法的影响

Influence of Reference Points on WiFi Location Fingerprinting Algorithm

随着室内定位技术的发展,WiFi位置指纹定位算法得到了人们的广泛关注,主要探讨了参考点数量与参考点间隔对于算法精度的影响。采用Matlab软件对位置指纹算法分别在实验环境和模拟环境中进行了仿真,通过实验数据得出随着参考点数量的增加,KNN与WKNN算法定位精度差不断减小,参考点数量越多,后改进的WKNN算法相比于KNN算法的定位精度提高的也越来越小,基本无明显差别。参考点数量越多计算越复杂,采用白化的K-Means聚类算法提高计算效率和定位精度,对参考点间隔对于定位精度的影响进行了实验,得出定位精度并不是参考点间隔越小越精确,而是在1.5 m左右的间隔为佳。

With the development of indoor positioning technology, WiFi location fingerprinting algorithm has attracted much attention. The influences of the number of reference points and reference point intervals on the accuracy of the algorithm are mainly discussed. The location fingerprinting algorithm was simulated respectively in the experimental environment and the simulation environment by Matlab software. The experimental data showed that with the increase of the number of reference points, the difference in positioning accuracy between KNN and WKNN algorithms was decreasing. Compared with the KNN algorithm, the more the number of reference points are, the smaller the location accuracy of the improved WKNN algorithm is, and there is no obvious difference basically. The more the number of reference points, the more complex the calculation. Whitening KMeans clustering algorithm was used to improve the computational efficiency and positioning accuracy, the effect of reference point intervals on the positioning accuracy was experimented. It is concluded that the positioning accuracy is not as precise as reference point interval is smaller, but better at about 1.5 m interval.