High-Accuracy Indoor Localization Utilizing Rf Wireless Location Signatures And High-Fidelity Predictive Models

This paper presents theoretical and experimental location accuracy results for locating mobile cellular phones in dense-urban indoor environments utilizing the 3GPP-standardized Radio Frequency Pattern Matching (RFPM) method and employing new high-fidelity predictive RF Models. An extensive field experiment involving thousands of distinct spatially-diverse test points is described; this test campaign covered a large number of representative dense-urban indoor and outdoor environments. The location results utilize as data input only cellular network measurements made by the mobile device as part of its normal operation for mobility management; these measurements are available on existing standard cellular network interfaces. The location system employs Wireless Location Signatures (WLSTM), which is Polaris Wireless's implementation of 3GPP-standardized RFPM; WLS has been widely deployed domestically and internationally in dozens of markets and carrier networks. This article describes new features of high-fidelity predictive models for building accurate RF pattern databases, one of the important key inputs in RFPM. Finally, the experimental results show that the location accuracy of RFPM in dense-urban and indoor environments can perform better than the FCC's location accuracy mandate for network-based E911 solutions; namely an indoor handset can be located to better than 80 meters 67 percent of the time and to better than 250 meters 90 percent of the time.