Date of Award
Master of Science (MS)
Indoor positioning systems (Wireless localization), Ubiquitous computing, Radio frequency
Computation of accurate indoor positioning information is important in several application areas like mobile robotics, navigation, virtual reality and ubiquitous computing. In spite of its importance we still do not have an indoor positioning system equivalent to the Global Positioning System~(GPS), which is the standard for large scale outdoor positioning, but is unreliable or unavailable in indoor environments. In this work, we present a multi-source radio frequency(RF) based system for automatic indoor positioning using the received signal strength (RSS) and demonstrate its efficacy by simultaneously using broadcast FM radio & GSM signals for position estimation. This method is data driven and can be justified using a Bayesian minimum risk framework. The methods used in this work is easy to extend by incorporating other sources of RF signals (like AM/Wi-Fi radio signals) and thus provides a generalized framework for building an indoor positioning systems using signals of opportunity. Using our algorithms and the well known AMBILOC dataset, we can localize exactly for approximately 98.7 % of the test locations in indoor settings over different months in a year, which demonstrates not only the efficacy of the algorithms but also its resiliency to change in the RF environment across the year and hence shows the transfer learning capability of our system.
Perekadan, Vishal Chummar, "Indoor positioning using multiple signal sources (modes)" (2019). Theses. 287.