Wireless Indoor Localization Via Crowdsensing

Location contexts act as a primary element of mobile internet and a fundamental type of sensing data for Internet of Things and thus plays a crucial role in our everyday life. People spend more than 80% of their time indoors, stimulating the flourishing markets and increasing demand of Location-based Services(LBS). During the past decades, various indoor location technologies have been proposed. Wi Fi-based localization for smartphones, among others, becomes attractive and popular, thanks to the wide availability of wireless networks as well as the global popularization of smartphones. A predominant location system that provides worldwide and round-the-clock services, however, still lacks. Even the mainstream Wi Fi fingerprint-based scheme encounters critical challenges in practice. Particularly, its applicability is extremely limited by severe drawbacks including high deployment costs due to the cumbersome site survey, biased radio maps caused by environmental dynamics especially during long-term running, and unsatisfactory performance in mobile environments, etc. This dissertation targets at the above problems and studies crowdsensing-based wireless indoor localization by designing a high-accuracy system at low hardware costs, low deployment costs and low maintenance costs, which enables wireless indoor localization deployable at large scale, runnable over long term, and applicable with high accuracy. In summary, the major results and core contributions are as follows:(1) Automatic construction of radio maps. Leveraging crowdsensed user mobility, we design an automatic method for radio map construction method, which eliminates the efforts of time-consuming and labor-intensive site survey by the participatory sensing of a number of mobile users. The proposed scheme significantly decreases the deployment costs of fingerprint-based localization. Results from real experiments demonstrate that comparable performance can be achieved by automatically built radio maps compared with manually calibrated ones.(2) Adaptable self-updating of radio maps. An automatic and continuous radio map self-updating service is proposed for wireless indoor localization, which exploits the static behaviors of mobile devices. By accurately pinpointing mobile devices with a novel trajectory matching algorithm, we employ them as mobile reference points to collect real-time RSS samples when they are static. With these fresh refer-ence data, the complete radio map is adapted by learning an underlying relationship of RSS dependency between nerghboring locations, which is expected to be relatively constant over time. The effectiveness is validated through real experiments,which show 2× improvement in location accuracy by appropriately adapting the radio maps.(3) An accurate localization algorithm for mobile environments. This dissertation investigates fingerprint-based localization in mobile environments and reveals crucial observations that act as the root causes of location errors, yet are surprisingly overlooked or not adequately addressed previously. Specifically, we recognize APs’ diverse discrimination for fingerprinting a specific location, observe the RSS inconsistency caused by signal fluctuations and human body blockages, and uncover the transitional fingerprint problem on commodity smartphones. Inspired by these insights, we devise a unified scheme of fingerprint generation, representation, and matching for mobile devices, which defines a discrimination factor to quantify different APs’ discrimination, incorporates robust regression to tolerate outlier measurements, and reassembles different normal fingerprints to cope with transitional fingerprints. Extensive experiments validate that the location errors are decreased by more than 50% compared to several state-of-the-art schemes.