Vision-Assisted Wireless Peer-to-Peer Indoor Localization

With the rapid development of mobile communications in recent years,the number of smart device users has exploded.Recent years have witnessed great advances in Internet of things.At the same time,the user's demand for location is increasing.Indoor localization has been widely used in hospitals,large shopping malls and other areas,providing convenience and safety for users.However,the existing indoor localization systems often need fingerprinting in advance or require a large amount of localization infrastructure such as access points.These prerequisites limit the application of these systems in searching for lost digital devices or finding lost children in unknown environments.To this end,we proposed a peer-to-peer localization system which is able to exploit one on-body device to directly locate another device without fingerprinting or localization infrastructure.Peer-to-peer localization can be divided into two categories,static targets localization and moving targets localization.We designed peer-to-peer localization method based on wireless and visual information obtained from the mobile devices.For the static target,we designed a vision-assisted antenna array construction scheme and radio peer-to-peer localization system,which exploited one-dimensional wireless information(Ao A).Based on the idea of Synthetic Aperture Radar(SAR),we leveraged the camera mounted on smartphone to achieve high-precision movement trajectory recovery,then we exploited the wireless signal of each point on the movement trajectory to realize an accurate virtual antenna array.To solve the phase errors of Channel State Information(CSI)caused by hardware defects in commercial equipment,we proposed an amplitude-based angle of arrival estimating method to achieve high-precision direction information.Finally,by combining the nonlinear characteristics of the moving trajectory and Ao A estimation,we used triangle localization to acquire the location of the static target.For the moving target,based on the vision-assisted antenna array,we designed a peer-to-peer localization scheme using multi-dimensional wireless information such as Time of Flight(To F),doppler frequency shift,and Ao A.First,we designed a direct path To F estimation method,then put forward a distance estimation scheme that combines Doppler frequency shift,so as to obtain high-precision distance information.Finally,we fused this distance and the above-mentioned Ao A to acquire the location of moving targets by using particle Swarm Optimization and Global Search.We implement our algorithm by using two smartphones,and the results verify the effectiveness of the indoor peer-to-peer localization system proposed in this thesis.