一种基于WKNN定位的改进算法

An Improved Algorithm based on WKNN Positioning

指纹定位算法是一种基于RSSI的定位算法。常见的指纹定位算法包括NN、KNN、WKNN。其中,WKNN是带有权重的K最近邻法,依据每个样本点对未知节点的贡献程度给每个指纹赋予一个权值。由现有文献可知,WKNN的定位精度优于NN、KNN,但是仍然存在定位精度有限的问题。为进一步提高WKNN算法的定位精度,减少定位误差,提出一种基于WKNN定位的改进算法。改进算法的思路是在WKNN算法的基础上结合极大似然估计算法。在MATLAB平台下进行仿真,仿真结果表明:在相同的仿真环境下,改进算法的定位精度明显高于现有的WKNN算法,定位误差明显小于WKNN算法。

Fingerprint localization algorithm is a localization algorithm based on RSSI. Common fingerprint localization algorithms include NN,KNN and WKNN. Among them,WKNN is K-nearest neighbor method with weight,which gives each fingerprint a weight according to the contribution of each sample point to the unknown node. According to the existing literature,WKNN has better positioning accuracy than NN and KNN,but there are still some problems with limited positioning accuracy. To further improve the positioning accuracy of WKNN algorithm and reduce positioning error,an improved algorithm based on WKNN positioning is proposed. The idea of improved algorithm is to use the maximum likelihood estimation algorithm after WKNN algorithm. Simulated in MATLAB platform,the results show that in the same environment,the localization accuracy of the improved algorithm is significantly higher than that of the existing WKNN algorithm,and the positioning error is significantly smaller than the WKNN algorithm.