Exploiting auxiliary information for indoor smartphone user tracking

Extensive deployment of wireless infrastructure provides an alrernative ability to locate smart phones in indoor environments by the use of received signal strength (RSS). This low-cost technology is, however, susceptible to environmental distortions, requiring sophisticated signal processing. In this paper, we propose to design a Viterbi algorithm under a double-layer hidden Markov model (DHMM) for reliable smartphone user tracking. Use of a batch processing scheme enables the DHMM Viterbi algorithm to effectively make use of auxiliary information such as room topology and user orientation for estimating the most likely user location sequence (route). Comparisons between the proposed Viterbi algorithm and an existing filtering algorithm from a theoretical perspective are also conducted. Experimental results show that use of the Viterbi algorithm is able to provide more reliable indoor positioning results.