Real-Time Machine Learning for Embedded Anomaly Detection

TL;DR

The paper tackles real-time, on-device anomaly detection in resource-constrained IoT/embedded systems. It surveys four main approach families—tree-based, one-class learning, lightweight neural networks, and statistical/threshold methods—and evaluates them through hardware- and latency-centric lenses. It highlights the dominant trade-offs, notes the rise of cascaded/hybrid designs, and identifies critical gaps in benchmarking, drift adaptation, adversarial robustness, and cross-platform portability. It concludes with practical guidance and a roadmap for hardware-aware, TinyML deployments of edge anomaly detection.

Abstract

The spread of a resource-constrained Internet of Things (IoT) environment and embedded devices has put pressure on the real-time detection of anomalies occurring at the edge. This survey presents an overview of machine-learning methods aimed specifically at on-device anomaly detection with extremely strict constraints for latency, memory, and power consumption. Lightweight algorithms such as Isolation Forest, One-Class SVM, recurrent architectures, and statistical techniques are compared here according to the realities of embedded implementation. Our survey brings out significant trade-offs of accuracy and computational efficiency of detection, as well as how hardware constraints end up fundamentally redefining algorithm choice. The survey is completed with a set of practical recommendations on the choice of the algorithm depending on the equipment profiles and new trends in TinyML, which can help close the gap between detection capabilities and embedded reality. The paper serves as a strategic roadmap for engineers deploying anomaly detection in edge environments that are constrained by bandwidth and may be safety-critical.