闽江学院学报

针对超宽带(UWB)定位精度受多径效应和非视距(NLOS)等因素影响的问题，分析了HDS-TWR方法的测距误差特征，提出了一种基于基站筛选和K-Means融合的UWB室内定位算法。首先，引入拉依达准则对测距数据中的异常值或缺失值进行处理，其次，设计了基站筛选和最小二乘法相结合的标签初始定位算法，最后，基于K-Means聚类算法确定k个可信度高的参考位置及聚类点密度，进而利用基于密度的加权质心算法求出标签的最优位置。实验结果表明，对比Chan算法和LSM-Kalman算法和Chan-KMeans算法等，该算法定位精度分别提高了48.86%、23.29%和41.60%。

Aiming at the issue of the impact of multipath effects and non line of sight(NLOS) factors on the positioning accuracy of ultra wideband(UWB).Based on the error characteristics analyses of HDS-TWR ranging, a UWB indoor positioning algorithm integrate K-Means and base station selective mechanism was proposed.Firstly, the Pareto criterion is adopted to deal with the outliers or missing values of the distance measurement data(between the base station and the label).Secondly, and an initial positioning algorithm was proposed based on the LSM with base station selective mechanism.Finally, the K-Means clustering algorithm was adopted to determine the top k reliable reference positions, as well as the densities, then the optimal position was determined with the density based weighted centroid algorithm.The experimental results demonstrated that the positioning accuracy of the proposed algorithm has been improved by 48.86%,23.29%,and 41.60% respectively,compared with Chan algorithm,LSMKalman algorithm,and Chan K-Means algorithm.