基于WFCM算法的室内指纹定位算法

Indoor fingerprinting localization algorithm based on weighted fuzzy C-means clustering algorithm

提出一种指纹定位的改进算法,利用RSSI和WFCM算法,通过对参考节点和未知节点进行聚类来实现定位。改进算法不需要未知节点与所有参考节点进行相似度比较,只需与所在同一类别的参考节点比较即可,克服了指纹定位耗时耗力和三角定位精确度低的问题。同时,采用欧几里得距离作为比较相似度的参数,同时定位仿真过程中取不同的噪声误差。仿真结果表明,与FCM算法相比,该算法能更好地实现聚类,定位更加精确,算法的鲁棒性更高。

This paper proposes an improved fingerprinting positioning algorithm based on the received signal strength （RS- SI） and weighted fuzzy C-means clustering （WFCM） using reference nodes and unknown nodes to achieve accurate positioning results. The improved algorithm does not require similarity comparison of the unknown nodes with all of the reference nodes and the unknown nodes are simply compared with the reference nodes of the same class, which can overcome the time-consuming problem of the fingerprinting localization and the low accuracy problem of triangular positioning. Meanwhile, Euclidean distance is used as the similarity comparison parameter and noise errors of different level are taken during the simulation. Simulation re-suits show that compared with the fuzzy C-means clustering （FCM） algorithm, the proposed algorithm can achieve better cluster results and higher positioning accuracy. And the algorithm also has higher robustness.