CAJ | 학술논문

阐述了位置指纹定位算法在室内WIAN环境中的应用，分析了KNN定位算法存在的不足，提出一种模糊聚类KNN位置指纹定位算法。该算法首先选取与空间相关性较好的4个信号参数，构成多径纹信号数据库；然后应用主分量分析法（PCA）对原始信号数据库作降维运算，滤除奇异性接入点（AP）；最后用模糊C均值聚类算法（FCM）处理数据，进一步滤除奇异性参考点（RP），实现提高定位算法效率与精度的目的。实验表明，改进后的定位算法产生的定位误差明显减小。
천술료위치지문정위산법재실내WIAN배경중적응용，분석료KNN정위산법존재적불족，제출일충모호취류KNN위치지문정위산법。해산법수선선취여공간상관성교호적4개신호삼수，구성다경문신호수거고；연후응용주분량분석법（PCA）대원시신호수거고작강유운산，려제기이성접입점（AP）；최후용모호C균치취류산법（FCM）처리수거，진일보려제기이성삼고점（RP），실현제고정위산법효솔여정도적목적。실험표명，개진후적정위산법산생적정위오차명현감소。
The paper describes the application of location fingerprint positioning algorithm in indoor WLAN environment, and analyzes the shorteomings of the KNN positioning algorithm, and proposes a fuzzy clustering KNN location fingerprint positioning algorithm. Firstly, it selects ibur signal parameters with good spatial correlation, and constitutes a database of muhipath pattern signal. Secondly, it does dimensionality reduction operations to the original signal database by application of principal component analysi(PCA), to filter out the the singularity access point(AP). Then, by application of fuzzy C-means clustering algorithm(FCM), it processes the data to further filter the singular reference point (RP), to achieve the purpose of improving the efficiency and accuracy of the positioning algorithm. The experiments show that the positioning errors generated by the improved positioning algorithm is significantly reduced.