改进K均值聚类BP神经网络超宽带室内定位方法

Improved K-means Clustering BP Neural Network UWB Indoor Location Method

室内定位技术随着生产生活对室内定位的需求不断提高而成为热点问题之一,超宽带定位技术以其独特的优势而快速发展,针对超宽带定位系统定位数据存在稳定性差、精度低的问题,提出改进K均值聚类BP神经网络超宽带室内定位方法。该方法对超宽带定位系统中原始数据进行改进K均值聚类算法的预处理,过滤位置偏差大的数据。再对测定区域内多点进行基于真实坐标和聚类过滤坐标的BP神经网络建模。定位过程则将过滤后的待定位坐标输入建好的BP神经网络定位模型,进行识别定位。通过实验测定,该室内定位方法精度可达26cm,且稳定性好。

Indoor positioning technology with the production and living on the indoor positioning of the demand has become one of the hot issues, ul- tra-wideband positioning technology with its unique advantages and rapid development, for ultra-wideband positioning system positioning data is poor stability, low accuracy, proposes the improved K-means clustering BP neural network UWB indoor location method. This meth- od improves the original data in the UWB positioning system and preprocesses the K-means clustering algorithm to filter the data with large deviation of the position. And then carries on the BP neural network modeling based on the real coordinates and the clustering filter coordinates in the multi-point in the measurement area. Positioning process is filtered after the positioning of the coordinates of the input BP neural network positioning model, identification and positioning. Through the experimental determination, the indoor positioning meth- od accuracy of up to 26cm, and good stability.