基于对数正态模型的二次匹配指纹定位算法

Second Matching Fingerprint Positioning Algorithm Based on Log-Normal Model

针对复杂室内环境WiFi信号不稳定导致指纹定位算法定位精度不高的问题,对AP选择、指纹库构建和定位算法进行研究,分别提出了基于对数正态模型的AP选择方法、RSSI虚拟指纹库构建方法以及基于对数正态模型的二次匹配指纹定位算法。在离线阶段,通过拟合对数正态模型对WiFi信号建模,筛选出最契合模型的AP,并构建RSSI指纹库和虚拟指纹库。在线定位阶段,待定位节点在匹配RSSI指纹库的基础上,利用AP的契合度修正欧氏距离相似度,找出相似度更高的指纹点,初步确定待定位节点区域,接着二次匹配虚拟指纹库进行精准定位。实际环境实验结果表明,本文提出的方法突出了不同AP设备对定位造成的影响,能减少离线阶段指纹采集的工作量,有效的提高了定位精度。

Aiming at the problem that the positioning accuracy of fingerprint positioning algorithm is not high due to the instability of WiFi signal in complex indoor environment,this paper studies the AP selection,fingerprint library construction and positioning algorithm,and proposes an AP selection method based on log-normal model,A method for constructing the RSSI virtual fingerprint library,and the second matching fingerprint positioning algorithm based on log-normal model.In the offline phase,by fitting the log-normal model,the WiFi signal is modeled,and the AP that fit the model are selected to construct the RSSI fingerprint library and the virtual fingerprint library.In the online positioning stage,on the basis of matching the RSSI fingerprint library,the unknow node uses the fit degree of AP to correct the Euclidean distance similarity,finds the node with higher similarity,determines the area initially,and then matches the virtual fingerprint library again for precise positioning.The actual environmental experiment results show that the proposed method highlights the impact of different AP equipment on positioning,and reduces the workload of fingerprint collection in the offline phase,which can effectively improve the positioning accuracy.