Real-Time Adaptive Anomaly Detection in Industrial IoT Environments

TL;DR

This paper tackles real-time anomaly detection in dynamic Industrial IoT (IIoT) environments where data streams are high-velocity, high-dimensional, and non-stationary due to concept drift. It introduces SAPDAD, a scalable, drift-adaptive, prediction-driven framework that blends a multi-source LSTM predictor (PDAD-SID) with RealTimeOAW drift adaptation and GA-based hyperparameter optimization, augmented by PCA-based dimensionality reduction and seasonal feature decomposition. SAPDAD achieves superior accuracy and real-time performance across three real-world datasets, with AUCs up to $89.71\%$ and sub-millisecond per-record processing times, while maintaining scalability as dimensionality grows. The work demonstrates that combining real-time drift adaptation, multi-source predictions, and efficient feature reduction yields robust anomaly detection suitable for production IIoT settings, offering practical impact for reliability and self-healing industrial systems.

Abstract

To ensure reliability and service availability, next-generation networks are expected to rely on automated anomaly detection systems powered by advanced machine learning methods with the capability of handling multi-dimensional data. Such multi-dimensional, heterogeneous data occurs mostly in today's industrial Internet of Things (IIoT), where real-time detection of anomalies is critical to prevent impending failures and resolve them in a timely manner. However, existing anomaly detection methods often fall short of effectively coping with the complexity and dynamism of multi-dimensional data streams in IIoT. In this paper, we propose an adaptive method for detecting anomalies in IIoT streaming data utilizing a multi-source prediction model and concept drift adaptation. The proposed anomaly detection algorithm merges a prediction model into a novel drift adaptation method resulting in accurate and efficient anomaly detection that exhibits improved scalability. Our trace-driven evaluations indicate that the proposed method outperforms the state-of-the-art anomaly detection methods by achieving up to an 89.71% accuracy (in terms of Area under the Curve (AUC)) while meeting the given efficiency and scalability requirements.

Figures (10)

• Figure 1: Illustration of an IIoT-based smart manufacturing environment as the motivating scenario.
• Figure 2: System model.
• Figure 3: Overview of the proposed Scalable and Adaptable Prediction-Driven Anomaly Detection (SAPDAD).
• Figure 4: An example demonstrating the calculation of SID with the maximum prediction horizon length L=3
• Figure 5: An example of the sliding window and adaptive window of the proposed RealTimeOAW algorithm.