Abstract
Traditional radio-map-based localization methods need to sample a large number of location fingerprints offline, which requires huge amount of human and material resources. To solve the high sampling cost problem, an automatic radio-map construction algorithm based on crowdsourcing is proposed. The algorithm employs the crowd-sourced information provided by a large number of users when they are walking in the buildings as the source of location fingerprint data. Through the variation characteristics of users’ smartphone sensors, the indoor anchors (doors) are identified and their locations are regarded as reference positions of the whole radio-map. The AP-Cluster method is used to cluster the crowdsourced fingerprints to acquire the representative fingerprints. According to the reference positions and the similarity between fingerprints, the representative fingerprints are linked to their corresponding physical locations and the radio-map is generated. Experimental results demonstrate that the proposed algorithm reduces the cost of fingerprint sampling and radio-map construction and guarantees the localization accuracy. The proposed method does not require users’ explicit participation, which effectively solves the resource-consumption problem when a location fingerprint database is established.
Highlights
Wireless sensor networks (WSNs), which are multi-hop self-organized networks with wireless communication, are used to achieve information collection and processing by deploying a large number of tiny sensor nodes in a surveillance area
As for the complex indoor environments, static or mobile users can be located by the received signal strength (RSS) [3], time of arrival (TOA) [4], time difference of arrival (TDOA) [5] and angle of arrival (AOA) [6]
Since the number of fingerprint clusters directly determines the size of the radio-map, the bigger the radio-map is, the fingerprint clusters directly determines the size of the radio-map, the bigger the radio-map is, the more more reference information it provides for online localization
Summary
Wireless sensor networks (WSNs), which are multi-hop self-organized networks with wireless communication, are used to achieve information collection and processing by deploying a large number of tiny sensor nodes in a surveillance area. Each piece of Wifi information and its corresponding location coordinates constitute a location fingerprint, and all the location fingerprints constitute a radio-map of the indoor area. If all smart phones carried by ordinary users implicitly take part in the construction of a radio-map and contribute data to the acquisition of indoor fingerprints, the sampling workload can be significantly reduced. In order to solve the poor extensibility and flexibility of radio-map construction in traditional fingerprint-based indoor localization algorithms, a novel automatic radio-map construction algorithm based on crowdsourcing (RACC) is proposed in this paper. Extensive simulations show that RACC can automatically generate radio-maps with acceptable accuracy and efficiently solve the resource-consuming problem in traditional radio-map-based indoor localization methods. The remainder of the paper is organized as follows: Section 2 discusses two categories of methods in radio-map-based indoor localization algorithms.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
