CAJ | 학술논문

位置指纹法定位中所建立的位置指纹库能否正确反映指纹信号与位置之间的映射关系将影响最终定位效果。本文将一种改进的K-means聚类算法运用于WLAN室内指纹数据库的建立，较之传统K-means聚类算法所建立的指纹数据库，采用了改进K-means聚类算法所建立起来的指纹数据库优化了初始聚类中心选择方法及准则函数，避免了被聚类的指纹数据陷入局部解现象。实验结果表明：本文所提算法建立的指纹数据库可提高WLAN室内定位精度，缩短定位时间。
위치지문법정위중소건립적위치지문고능부정학반영지문신호여위치지간적영사관계장영향최종정위효과。본문장일충개진적K-means취류산법운용우WLAN실내지문수거고적건립，교지전통K-means취류산법소건립적지문수거고，채용료개진K-means취류산법소건립기래적지문수거고우화료초시취류중심선택방법급준칙함수，피면료피취류적지문수거함입국부해현상。실험결과표명：본문소제산법건립적지문수거고가제고WLAN실내정위정도，축단정위시간。
Whether positioning fingerprint database is created in location fingerprinting position can accurately reflect the mapping relationship between the position and the fingerprints signal and it will affect the final precise of positioning. In this paper, an improved K-means clustering algorithm is used to build WLAN indoor fingerprint data-base. Compared with the fingerprint database built by traditional K-means clustering algorithm, the improved algo-rithm enables the database to avoid being trapped in local optimum by the application on the optimization of the method of the selection of initial database cluster center and the criteria function. Experimental results show that pro-posed algorithm which yield fingerprint database improves the accuracy of WLAN indoor positioning and shorts the running time.