摘要
Abstract
Current localization methods on social platforms predominantly rely on distance mapping,which necessitates extensive queries to identify critical thresholds.However,as platforms continually strengthen their location obfuscation mechanisms,distance-based approaches struggle to accurately characterize the perturbation patterns,resulting in a significant drop in localization accuracy.A Soul user geolocation method based on direction constraint and double fan-shaped ring intersection is pro‑posed.First,an omni-directional and multi-range virtual probe deployment strategy is designed,which adaptively sets multi-level radial sampling radii according to the display granularity of platform dis‑tance,thereby efficiently acquiring non-redundant mapping data.On this basis,the azimuth is incorpo‑rated into location modeling to construct a location bandwidth model considering both distance and azi‑muth.Fan-shaped annular regions are generated from two distinct observation points,and the geomet‑ric centroid of their intersection is used as the final user location estimate.Experiments on Soul show the minimum,average and maximum localization errors of 9.01,29.87 and 57.66 m,respectively.Com‑pared with 3 typical methods,the maximum reductions of the 3 aforementioned errors using the pro‑posed method can reach 71.78%,57.12%and 43.16%,respectively.关键词
基于位置的社交平台/用户定位/位置映射模型/方位角Key words
location-based social platform/user geolocation/location mapping model/azimuth分类
信息技术与安全科学
