Decision Fusion for IoT-Based Wireless Sensor Networks

This article presents a novel decision fusion algorithm for Internet-of-Things-based wireless sensor networks, where multiple sensors transmit their decisions about a certain phenomenon to a remote fusion center (FC) over a wide area network. The proposed algorithm denoted as the individual likeliho...

Full description

Saved in:
Bibliographic Details
Published in:IEEE internet of things journal 2020-02, Vol.7 (2), p.1313-1326
Main Authors: Al-Jarrah, Mohammad A., Yaseen, Maysa A., Al-Dweik, Arafat, Dobre, Octavia A., Alsusa, Emad
Format: Article
Language:English
Subjects:
Algorithms
Cloud computing
Codes
Computer simulation
Data fusion
Error analysis
global connectivity
Internet of Things
Internet of Things (IoT)
long-range wide area network (LoRaWAN)
Monte Carlo simulation
narrowband IoT (NB-IoT)
Performance degradation
Probability
Reliability
Remote sensors
Sensor fusion
Sensor phenomena and characterization
Transition probabilities
Wide area networks
Wireless networks
wireless sensor network (WSN)
Wireless sensor networks
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This article presents a novel decision fusion algorithm for Internet-of-Things-based wireless sensor networks, where multiple sensors transmit their decisions about a certain phenomenon to a remote fusion center (FC) over a wide area network. The proposed algorithm denoted as the individual likelihood approximation (ILA) can significantly reduce the decision fusion error probability performance while maintaining the low computational complexity of other state-of-the-art fusion algorithms. The performance of the ILA rule is evaluated in terms of the global fusion probability of error, and an efficient analytical expression is derived in terms of a single integral. The analytical results corroborated by Monte Carlo simulation show that the ILA significantly outperforms all other considered rules, such as the Chair-Varshney (CV) and MaxLog rules. Moreover, the impact of the link from the cluster head to the FC, which is modeled as a binary symmetric channel with unknown transition probabilities, has been investigated. It is shown that the probability of error over such links should not exceed 10 -3 to avoid severe performance degradation. Furthermore, we derive a closed-form expression for the system fusion error probability of the CV rule for the most general system parameters.
ISSN:2327-4662
2327-4662
DOI:10.1109/JIOT.2019.2954720