基于动态时间规整距离指纹匹配的Wi-Fi网络室内定位算法

发布时间：2018-04-04 15:43

本文选题：Wi-Fi网络　切入点：室内定位　出处：《计算机应用》2017年06期

【摘要】：Wi-Fi网络中常规的基于指纹匹配室内定位算法面临信号时变现象或人为干扰的影响,导致定位精度不高。为此,提出基于动态时间规整(DTW)距离相似性指纹匹配的Wi-Fi网络室内定位算法。首先,该算法将定位区域的Wi-Fi信号特征按照采样的先后顺序转化为时间序列类型指纹,通过计算Wi-Fi信号指纹动态时间规整距离的大小来获取定位点与样本点的相似性;然后,根据采样区域结构特征,将Wi-Fi信号指纹采集问题划分为三类基本的动态路径采样方式;最后,结合多种动态路径采样方式增加指纹特征信息的准确性和完整性,从而提高指纹匹配的准确性和定位精度。大量实验结果表明,较瞬时指纹匹配定位算法,所提算法误差范围在3m以内定位的累积错误率:路径区域匀速运动提高了10%,变速运动提高了13%;开放区域交叉曲线运动提高了9%,S型曲线运动提高了3%。所提算法在实际室内定位应用中能有效提高指纹匹配的准确性和定位精度。
[Abstract]:The conventional fingerprint matching indoor location algorithm in Wi-Fi network faces the influence of signal time-varying phenomenon or artificial interference, which leads to the low accuracy of location.Therefore, an indoor location algorithm for Wi-Fi network based on dynamic time warping (DTW) distance similarity fingerprint matching is proposed.Firstly, the Wi-Fi signal feature of the location region is transformed into time series fingerprint according to the sequence of sampling, and the similarity between the location point and the sample point is obtained by calculating the dynamic time warping distance between the Wi-Fi signal fingerprint and the sample point.According to the structure of the sampling region, the fingerprint acquisition problem of Wi-Fi signal is divided into three basic dynamic path sampling methods. Finally, the accuracy and integrity of fingerprint feature information are increased by combining various dynamic path sampling methods.In order to improve the accuracy of fingerprint matching and positioning accuracy.A large number of experimental results show that, compared with the instantaneous fingerprint matching algorithm,The error range of the proposed algorithm is less than 3 m. The cumulative error rate of the proposed algorithm is as follows: the uniform motion of the path region is increased by 10 percent, the movement of the variable speed increases by 13 percent, and the movement of the cross curve in the open area increases by 9 percent and the movement of the S-shaped curve increases by 3 percent.The proposed algorithm can effectively improve the accuracy and accuracy of fingerprint matching in practical indoor localization applications.
【作者单位】：东北大学计算机科学与工程学院;东软公司软件架构新技术国家重点实验室;
【基金】：国家863计划项目(2015AA016005) 国家自然科学基金资助项目(61402096,61173153,61300196)~~
【分类号】：TN92

【相似文献】