PEER-TO-PEER LOCALIZATION IN URBAN AND INDOOR ENVIRONMENTS

Abstract

This paper presents a novel peer-to-peer or mobile- to-mobile localization scheme for general indoor and outdoor environments. In this scheme, two mobile nodes at arbitrary locations are able to locate each other without the need of Line-of-Sight (LOS) path between the two mobile device, and without the need for any reference devices such as GPS or land base beacons. Existing peer- to-peer localization techniques make use of Time of Arrival (TOA) and Angle of Arrival (AOA) of LOS and single bounce scattering paths to derive line of possible mobile device positions (LPMDs). The intersections of LPMDs are then used to estimate the unknown mobile device position | referred to as the Line Segment Intersection. However, in a heavy multipath environment with many multiple- bounce scattering paths, existing techniques require weighting factors and threshold values which are speciflcally chosen for that particular environment in order to select the LPMDs that correspond to LOS and single-bounce scattering paths for localization. Large localization error will occur if multiple-bounce scattering paths' LPMDs are mistakenly used for intersections. In addition, existing techniques also do not work well in a multipath environment with high level of TOA and AOA noises especially when the angles between LPMDs are small. The accuracy of the Line Segment Intersection also deteriorates as the distance traveled by multipath signals become comparable to each other. This renders the weighting and threshold values inefiective. This paper presents a novel Gaussian weighting process to remove the abovementioned limitations. The Gaussian weighting process also dramatically improves the accuracy of the localization. Experimental coupled with simulation results show that our proposed localization scheme outperforms existing Peer-to-peer localization technique by a signiflcant margin of up to 83% and 54%