Machine learning-based fault detection scheme for IoT-enabled WSNs
by Pravindra Shekhar Shakunt; Siba K. Udgata
International Journal of Sensor Networks (IJSNET), Vol. 46, No. 2, 2024

Abstract: Wireless sensor networks (WSNs) encounter faults due to their deployment in non-deterministic and potentially hazardous environments. The process of fault identification in WSNs poses several challenges. This complexity makes it challenging to pinpoint and diagnose faults within IoT-enabled WSNs. This paper uses extra tree and state-of-the-art machine learning classifiers to classify commonly occurring faults such as offset, drift, gain, data loss, stuck, random, and out-of-bounds at the sensor node level. First, we realistically induce the faults with different intensities to a benchmark dataset of temperature and humidity sensors. We propose sliding window-based data pre-processing techniques and various machine learning algorithms for classifying types of faults. The performance of the proposed scheme and other machine learning approaches are compared based on specificity, precision, recall, accuracy, F1-score, and AUC-ROC performance evaluation metric. Experimental study shows that our proposed scheme and the extra tree machine learning approach are more effective than state-of-the-art approaches.

Online publication date: Tue, 01-Oct-2024

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Sensor Networks (IJSNET):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com