基于方差滤波的动态自适应室内指纹定位算法-计算机应用研究.pdfVIP

优先出版 计 算 机 应 用 研 究 第33 卷 基于方差滤波的动态自适应室内指纹定位算法 张力仁，沈建京 (信息工程大学 理学院，郑州 450001) 摘 要：目前常用的室内定位算法在构建指纹库时，因为在参考点处接收到的信号强度值存在噪声值，传统的欧式距离 计算公式忽略信号发射器的稳定程度从而造成定位误差，以及 K 近邻算法无法根据实际情况动态的确定参与定位的参 考点个数，从而造成定位算法缺乏拓展性和稳定性。针对上述情况，本文设计了一种经过方差滤波的动态自适应室内指 纹定位算法。该算法首先在离线阶段利用方差滤波去掉噪声值，之后在在线阶段利用新的基于方差的相似度计算公式计 算待定位点与各个参考点之间的相似度，然后动态自适应的确定距离较小的K 个点来参与定位计算，最后利用确定的K 个点加权估计待定位点的位置。实验结果证明，该算法较对比算法能够有效的提高定位精度和定位的稳定性。 关键词：K 近邻分类算法；方差滤波；动态自适应；欧式距离；指纹定位算法 中图分类号：TP391.4 Dynamic and adaptive indoor fingerprint localization algorithm based on variance filtering Zhang Liren, Shen Jianjing (College of Science, Information Engineering University, Zhengzhou, China) Abstract: The presence of noise value at the reference point of the received signal strength value when building a fingerprint database and traditional Euclidean distance ignore the stability of the access points resulting positioning error and K-nearest neighbor algorithm can not be dynamically determined the number of reference points involved in locating based on the actual circumstances leading to the lack of stability and expansion. For the above situation, this paper presents a dynamic and adaptive fingerprint indoor location algorithm based on variance filtering. Firstly, the algorithm uses the variance filter to remove the noise value in the offline stage. Secondly, the algorithm uses a new formula based on the variance to calculate the distance between the respective reference point and the point to be positioned. Finally, use K reference points dynamically determined to weighted calculate the point to be located. Exp