Research On Wlan Indoor Localization Method Based On Subarea Division

With the development of wireless communication technology and pervasive computing,location-based services(LBS)have been widely used,and localization technology is the key to achieve LBS.In the outdoor environment,Global Positioning System(GPS)can achieve high accuracy and meet the requiements of most users for outdoor LBS.In the indoor environment,WLAN fingerprinting localization has become a research focus due to its wide distribution,low costs,easy measurement and no need of extra hardware.However,the complexity of indoor environment,heavy workload of fingerprints collection and high computational complexity restrict the application of WLAN localization technology.For the above problems,this thesis proposes a WLAN fingerprinting localization model based on subarea division and focuses on the construction and division of fingerprint database.The main work of the thesis includes:(1)By analyzing the problem of fingerprints collection and matching efficiency faced by WLAN fingerprinting localization,and combining existing methods,we propose a improved model of WLAN fin,gerprinting localization which mainly improves the construction and division of fingerprints database and AP optimization process.(2)For the problem of huge workload of site survey,a fingeprints database construction method based on particle swarm optimization-quasi newton(PSO-QN)is proposed.Virtual reference points are introduced in the offline phase.Only a few reference points are assigned,and the received signal strength indicator(RSSI)of these points are collected.The RSSI of virtual reference points are obtained through the signal propagation model to relieve the burden of site survey.Considering about the complexity of indoor environment,the initial reference points are divided into subareas according to indoor structure such as corridors and rooms.The log-distance path loss model are constructed for each region and the,parameters are calculated by PSO-QN algorithm.(3)For the problem of high computational complexity,the subarea division method based on improved fuzzy c-means(IFCM)algorithm and AP optimization method based on hierarchical clustering are proposed.According to the idea of subarea division,the between-within proportion(BWP)index is utilized to select the best clustering number for the division of fingerprints database by IFCM algorithm.The subarea division doesn't stop until the number of reference points in each subarea are less than the threshold.AP optimization is executed in each sub-region by hierarchical clustering to select the APs with strong discriminant ability to reduce the dimensions of fingerprints.In the online positioning phase,region selection is performed by nearest neighbor(NN)algorithm,and then the weighted k-nearest neighbor(WKNN)algorithm based on pearson correlation coefficient(PCC)is used to calculate the coordinates of targets.Finally,a WLAN indoor localization system is designed and implemented based on the above research,and the positioning effect is displayed using web and app.The system makes LBS application more extensive and location information more accurate.