• Home
  • Search
  • Employ the Machine Learning Models for immediate Predictive Diagnostics: Anomaly detection of Industrial IoT
  • https://doi.org/10.1109/gcwcn66157.2025.11448332Copy DOI Icon

Employ the Machine Learning Models for immediate Predictive Diagnostics: Anomaly detection of Industrial IoT

  • Nov 22, 2025
  • Jubert R Pagunuran +5 more
Show More
  • Abstract
  • Literature Map
  • References
  • Similar Papers
Abstract

Industrial Internet of Things (IIoT) has transformed plants and factories, facilitating real-time monitoring, automation, and informed decision-making. Nevertheless, in such complex data-rich environments, identification of the fault in a timely manner, detection of the anomaly, has become one of the most important challenges. Unresolved anomalies in IIoT systems can cause unscheduled maintenance, slippage in quality, or even catastrophic failure. The conventional rule-based models are not capable of accommodating the complicated dynamics and the non-linear dependencies that widely exist in industrial systems. In this study, the mixture machine learning (ML) models are applied for real-time predictive diagnostics and anomaly detection in the IIoT networks. System Overview The proposed system combines both supervised and unsupervised ML models getting the best of each of them such as Random Forest, Support Vector Machine (SVM), Autoencoders, Isolation Forest, Long Short-Term Memory (LSTM) networks. These models are learned and tested over massive time-series sensor data derived from IIoT ready industrial machines (e.g., pumps, motors, turbines, and conveyors). Method of feature extraction and the detection of temporal patterns are applied to find the abnormal behavior in the operational parameters: temperature, vibration, pressure and energy consumption. One of the key contributions in our study is that models could be deployed in real time and this is made possible by edge computation. By moving ML inference to the edge, the system provides quick response times even under bandwidth-constrained conditions. Alongside, long-term analytics trend forecasting and cross-facility model improvement is made possible via cloud-based aggregation. The combined alerting and strategic system along with hybrid structure allows operators to avoid catastrophic failures before they occur. Experiments on benchmark IIoT datasets show that the proposed models achieved excellent detection performance, e.g., more than 95% precision and recall for the temporal based model-LSTM. The ML-based approach is shown to vastly outperform conventional threshold-based detections in the presence of noisy, multivariate and asynchronous data. In addition, machine behavior evolves, and with it the ability to learn adapts diagnostic accuracy. This work underlines the disruptive nature of the ML for predictive maintenance and real-time industry diagnosis. Through a transition from reactive to proactive approaches, the developed framework minimizes downtime, lowers maintenance costs, and extends asset life. Incorporating smart diagnostics into the IIoT networks not only promotes operational efficiency, but also provides an infrastructural building block for the next-generation autonomous industrial systems that are data, intelligence, and connectivity enabled.

Similar Papers
  • Research Article
  • Citations16

Guest Editorial: Industrial IoT and Sensor Networks in 5G-and-Beyond Wireless Communication

  • Jun 01, 2022
  • IEEE Transactions on Industrial Informatics
  • Dong Yang +3
  • Dissertation
  • Citations3

Machine learning and deep learning-based anomaly detection for electric vehicle charging infrastructure and industrial internet of things

  • Jan 01, 2024
  • Sai Chandana Vallam Kondu
  • Research Article
  • Citations8

A novel deep learning framework with temporal attention convolutional networks for intrusion detection in IoT and IIoT networks

  • Dec 13, 2025
  • Scientific Reports
  • Kingkar Prosad Ghosh +4
  • Research Article
  • Citations2

How Should Enterprises Quantify and Analyze (Multi-Party) APT Cyber-Risk Exposure in their Industrial IoT Network?

  • Oct 19, 2023
  • ACM Transactions on Management Information Systems
  • Ranjan Pal +4
  • Research Article
  • Citations29

Cyberattack Detection Systems in Industrial Internet of Things (IIoT) Networks in Big Data Environments

  • Mar 13, 2025
  • Applied Sciences
  • Abdullah Orman
  • Conference Article
  • Citations7

Energy Consumption Optimization for UAV Assisted Private Blockchain-based IIoT Networks

  • Sep 01, 2021
  • Xinhua Lin +3
  • Research Article
  • Citations83

GA-mADAM-IIoT: A new lightweight threats detection in the industrial IoT via genetic algorithm with attention mechanism and LSTM on multivariate time series sensor data

  • Sep 04, 2024
  • Sensors International
  • Yakub Kayode Saheed +2
  • Research Article
  • Citations6

A Trustworthy Learning Technique for Securing Industrial Internet of Things Systems

  • Jan 01, 2021
  • Journal of Intelligent Systems and Internet of Things
  • Osama Maher +1
  • Research Article
  • Citations11

Trust-Based Certificate Management for Industrial IoT Networks

  • Jul 15, 2023
  • IEEE Internet of Things Journal
  • Chaimaa Boudagdigue +3
  • Research Article
  • Citations7

Massive MIMO for Massive Industrial Internet of Things Networks: Operation, Performance, and Challenges

  • Dec 01, 2024
  • IEEE Transactions on Cognitive Communications and Networking
  • Byung Moo Lee
  • Research Article

Wireless federated learning for salt-spray prediction in industrial IoT networks with delay constraint

  • Nov 23, 2025
  • Scientific Reports
  • Lincong Chen +6
  • Research Article

Enhance Capacity Region of Multi‐Class IIoT Network by Applying Entanglement Assistance Protocol

  • Jan 01, 2025
  • IET Quantum Communication
  • Doaa Subhi +1
  • Book Chapter
  • Citations14

Artificial Intelligence and Machine Learning for the Industrial Internet of Things (IIoT)

  • Feb 17, 2022
  • Fanoon Raheem +1
  • Dissertation

Spectrum: Towards Self-Aware Industrial IoT System

  • Oct 09, 2024
  • Sabari Nathan Anbalagan
  • Research Article

A Comprehensive Survey of Security Challenges and Solutions of Industrial Internet of Things (IIoT) Networks

  • Apr 24, 2024
  • INTERANTIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND MANAGEMENT
  • Anusha Kumari
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.