Autonomous Edge-Based Anomaly Detection in Industrial IoT

Abstract

ABSTRACT Industrial IoT (IIoT) systems are increasingly deployed in manufacturing, energy, and critical infrastructure to enable real-time monitoring and control. With the massive proliferation of IoT devices, detecting anomalies in system behavior has become critical to ensure reliability, safety, and operational efficiency. Traditional cloud-centric anomaly detection methods often suffer from latency, bandwidth constraints, and security concerns. Autonomous edge based anomaly detection emerges as a promising solution by processing data locally on edge devices with minimal reliance on central servers. This paper presents a comprehensive review of edge-based anomaly detection in IIoT, covering enabling technologies, architectures, algorithms, applications, and future research directions. Special attention is given to machine learning and deep learning approaches deployed on resource-constrained edge nodes. Challenges in scalability, real-time performance, and energy efficiency are also discussed. The study highlights the potential of autonomous edge intelligence to enhance IIoT resilience while reducing operational costs and network dependency.

KEYWORDS: Industrial IoT, Edge Computing, Anomaly Detection, Autonomous Systems, Machine Learning, Real-Time Monitoring, Predictive Maintenance