An intelligent multi-floor indoor positioning system for cloud-based environment. Issue 12 (2nd December 2022)

• Record Type: Journal Article
• Authors: Christy Jeba Malar, A.; Deva Priya, M.; Sengathir, J.; Kiruthiga, N.; Anitha, R.; Sangeetha, T.
• Is Part Of: International journal of computers and applications. Volume 44:Issue 12(2022)
• Page Start: 1170
• Page End: 1177
• Publication Date: 2022-12-02
• Subjects: Indoor node positioning -- wireless sensor network -- multi-floor clustering -- received signal strength; Computers -- Periodicals; Computer software -- Periodicals; Computer networks -- Periodicals; Multimedia systems -- Periodicals; Internet -- Periodicals; World Wide Web -- Periodicals; Minicomputers -- Periodicals; Microcomputers -- Periodicals; 004.05
• DOI: 10.1080/1206212X.2019.1696447
• Languages: English
• ISSNs: 1206-212X

Abstract

Abstract : Optimal node positioning to ensure energy efficiency is a challenge in Wireless Sensor Networks (WSNs). Energy can be conserved in WSN when the sensors are positioned at optimum locations. In this paper, an intelligent scheme to form multi-floor clusters by subdividing the topology into regions is proposed. Some random nodes are elected as Cluster Heads (CHs). The CH determines the number of neighboring nodes in each floor of the indoor surrounding. It sends requests to the members in the cluster to measure the Received Signal Strength (RSS). The node positions are estimated by applying Extreme Learning Machine (ELM) on the recorded RSS fingerprints. Kernelized ELM provides access to high dimensional input space with less computational cost and time. The proposed scheme provides more efficient control over the network during fluctuations. Once clusters are set up, the RSS of each node is collected to create a fingerprint database. The target node position is determined by matching the currently recorded RSS with the stored fingerprint. The performance of the proposed intelligent clustering technique is analyzed for both static and mobile nodes with varying samples and reference points. Extensive experiments conducted show that the propounded scheme offers better scalability and reduced Cumulative Percentile of Error (CPE).