| dc.description.abstract | Indoor positioning is a major challenge for location-based
services. WiFi deployments are often used to address indoor positioning.
Yet, they require multiple access points, which may not be available
or accessible for localization in all scenarios, or they make unrealistic
assumptions for practical deployments. In this paper we present
SPRING+, a positioning system that extracts and processes Channel
State Information (CSI) and Fine Time Measurements (FTM) from a
single Access Point (AP) to localize commercial smartphones. First, we
propose an adaptive method for estimating the Angle of Arrival (AOA)
from CSI that works on single packets and leverages information from
the estimated number of paths. Second, we present a new method
to detect the first path using FTM measurements, robust to multipath
scenarios. We evaluate SPRING+ in an extensive experimental campaign
consisting of four different testbeds: i) generic indoor spaces, ii)
generic indoor spaces with obstacles, iii) office environments and iv)
home environments. Our results show that SPRING+ is able to achieve
a median 2D positioning error between 1 and 1.8 meters with a single
WiFi AP. | es |
