Research On WLAN Outdoor Localization Method

In recent years,the rapid development of mobile internet and extensive use of smart terminals produce location-based service.Location-based service has been becoming a hot spot in industry application and plays an integral part of daily life.Typically,traditional positioning method like GPS could provide relatively good positioning service,but GPS consumes energy too quickly and mobile devices cannot maintain long battery life due to limited energy carrying capacity.Besides,in GPS positioning,higher positioning accuracy is based on good environmental conditions.Therefore,poor environmental conditions such as building occlusion may lead to large positioning errors.However,in indoor localization,Wi Fi-based fingerprinting localization is widely used and have gain good effects.Considering the widely deployment of Wi Fi infrastructure in urban area,it brings some new idea in outdoor localization.The research of this paper mainly focuses on utilizing Wi Fi infrastructure to make outdoor positioning in urban dense-constructed area,aiming to realize energy-saving purpose on the premise of ensuring a certain positioning accuracy.For the distribution of Wi Fi data in outdoor environment,this paper suggests to collect Wi Fi fingerprint data through crowdsourcing.After raw data cleaning and fingerprint database construction,distance-based algorithm calculates estimated location with the assistance of motion sensors in online localization phase.Experiments in the real environment fully proves that it is feasible to make positioning and save energy by utilizing Wi Fi information.In view of the high dimension problem in raw data and low accuracy of distancebased model,this research proposes a localization algorithm based on laplacian eigenmap.Specifically,by maintaining the structure in high dimension space and the adjacency relations among sample raw data,the estimated location is obtained by low-dimensional embedding in real location space.And in consideration of moving characteristics of mobile devices,we construct an optimization model based on condition constraints of sensor information and position time stability.The device location is estimated by solving this optimization model.Our research conducted real world experiments and verified our method,and it is mainly focused on the positioning accuracy,device power consumption rate and parameters sensitivity.The experimental results show that positioning accuracy of our method is slightly worse that the widely-adopted GPS but it has better power efficiency,indicating that our method is effective and has practical application significance.