Characteristics of crowdsourcing for outdoor radio fingerprinting positioning

Positioning systems based on location fingerprinting have become an area of intense research, mainly with the aim of providing indoor localization. Many challenges arise when trying to deploy location fingerprinting to an outdoor environment. The main problem is achieving coverage of large outdoor spaces, which costs time and needs an intensive data gathering effort. This paper proposes the use of mobile crowdsourcing in order to build a fingerprint database consisting of Wi-Fi networks received signal strength. The primary objective of this paper is to prove the feasibility of outdoor positioning using Wi-Fi crowdsourced data by showing that Wi-Fi networks are ubiquitous in urban areas. We then examine the gathered data and report our findings on challenges in building and maintaining a large-scale crowdsourced fingerprint database, the influence of fading on crowdsourced measurements. As Wi-Fi access-points are shown to exhibit mobility, we also propose and analyze methods for detecting and classification of mobile and static access-points.