Simultaneous pathway mapping and behavior understanding with crowdsourced sensing in WLAN environment

Abstract

With the wide deployment of Wireless Local Area Network (WLAN), the WLAN Received Signal Strength (RSS) based indoor mapping has attracted significant attention for various of Location-based Services (LBSs). In this paper, we propose a new simultaneous pathway mapping and behavior understanding approach by crowdsourced sensing of WLAN RSS. First of all, a batch of WLAN RSS sequences with known locations are collected to optimize the parameters in density clustering by using the concept of Jaccard Coefficient (JC). Secondly, the wavelet analysis is considered for each RSS sequence with unknown locations to mitigate the noise interference, and then the process of multidimensional scaling approach based dimension reduction is followed to reduce the computation cost involved in density clustering. Thirdly, based on the RSS and timestamp relations of RSS measurements, the density clustering is conducted to merge the RSS measurements at unknown locations into different clusters. Finally, the pathway mapping and behavior understanding are realized according to the indoor mapping from the signal into physical space. The extensive experimental results show that the proposed approach can achieve superior pathway mapping performance compared with the well-known WILL and CIMLoc approaches.