High-Accuracy Indoor Navigation Utilizing RF Wireless Location Signatures and High-Fidelity Predictive Models

J.J. Zhu, D. Qiu, J. Blaha, D.S. De Lorenzo, T. Bhattacharya

Abstract:	The growth of mobile navigation technologies has been enabled by the ubiquitous outdoor coverage of GPS and by the low cost of integrating a GPS radio into a small form-factor handheld device. This works well for outdoor environments, but when people move indoors – whether in malls, airports, or libraries – then mobile navigation performance degrades from the meter-level accuracy of current-generation (post-S/A) GPS to the tens or hundreds of meters of cellular location technologies. Recent advances in signal processing, high-fidelity predictive modeling, and multi-sensor fusion techniques now put us on the verge of near-ubiquitous high-accuracy indoor navigation to complement that available outdoors. The most promising technologies for indoor navigation include high-sensitivity GPS; Wi-Fi fingerprinting; Cell ID, cellular TOA, TDOA, and RF pattern-matching; and the integration of multiple sensors in conjunction with pedestrian dead-reckoning models for a true indoor navigation/way-finding experience. A great deal of academic and scholarly literature is available to describe many of these techniques, but there is a relative sparsity of material covering cellular RF pattern-matching for high-accuracy indoor location. Therefore, we describe in this paper an efficient and robust indoor location method based on 2G/3G/4G cellular signals, and focused on the specific use-case of rapid mobile location in order to provide E-911 emergency services. Particular limitations that apply to this application of indoor location include: a requirement to provide service to all mobile subscribers regardless of air-interface or handset capability, ensured rapid time-to-first-fix (<30 seconds) and 100% location yield, and FCC-mandated bounded accuracy. Furthermore, the data flow between the mobile device and the network must be supported by existing and evolving cellular telecommunications standards. Finally, there are significant and difficult-to-model attenuation, fading, scattering, and multipath effects that exist not only indoors but in any densely-built urban environment. We focus our attention in this study not only on theoretical analysis but also describe a variety of numerical simulation and experimental field-trial results. First, this paper describes the theory and methodology underlying RFPM technology, with a focus on theoretical performance in ideal conditions as well as degradation in non-ideal environments. The RFPM algorithms described in this paper employ high-fidelity radio propagation models to characterize the cellular RF environment as well as sophisticated estimation algorithms to build a robust prediction of where a mobile device is located based on the signals it observes; the specific implementation we describe is deployed in dozens of cellular networks both domestically and internationally. Second, a theoretical model based on handset measurement errors and database prediction errors in dense urban and indoor environment will be described. This theoretical model allows a prediction of the accuracy performance limit for an RFPM system in a variety of typical deployment environments. Third, an experimental field trial is described that exercises the RFPM algorithms in a representative dense urban and indoor environment. A model of the RF environment is constructed based on terrain and building databases, cellular network infrastructure and antenna coverage models, and high-fidelity RF propagation models. Verification of these models is conducted by vehicle and pedestrian methods. Finally, location accuracy results from these experimental field trials are presented. These results show that RFPM can do significantly better than the FCC’s location accuracy mandate for network-based solution; namely, we show that an indoor handset can be located to within 100 meters for 67 percent of the calls and to within 300 meters for 90 percent of the calls. We close this paper with some observations on limitations of these techniques and a description of ongoing and future research.
Published in:	Proceedings of the 26th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2013) September 16 - 20, 2013 Nashville Convention Center, Nashville, Tennessee Nashville, TN
Pages:	998 - 1005
Cite this article:	Zhu, J.J., Qiu, D., Blaha, J., De Lorenzo, D.S., Bhattacharya, T., "High-Accuracy Indoor Navigation Utilizing RF Wireless Location Signatures and High-Fidelity Predictive Models," Proceedings of the 26th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2013), Nashville, TN, September 2013, pp. 998-1005.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In