Tiny Machine-Learning Operations within Cyber-Physical Systems: a Field Study

2026-06-18

AI Summary

This work addresses the challenges of deploying TinyMLOps in safety-critical embedded systems, where model opacity, high data noise, and severe memory constraints hinder reliability. To overcome these limitations, the authors propose a knowledge-driven, end-to-end TinyMLOps framework that integrates physical principles, expert rules, and sensor data. The approach features automated time-series cleaning, knowledge-guided feature construction, interpretable regularized modeling, and rolling cross-validation to handle concept drift. Validated on industrial equipment, this method demonstrates for the first time that injecting domain knowledge significantly reduces false alarms and yields actionable decision rules. Implemented on an ARM Cortex-M4 microcontroller with only 32 kB of memory, the system achieves an AUC of 0.84, provides three-minute advance warnings of load peaks, cuts false positives by 50%, and reduces non-productive time by 11%, establishing a trustworthy engineering paradigm for TinyMLOps in cyber-physical systems.

Abstract

Machine-Learning Operations (MLOps) is maturing into a software-engineering discipline, yet its tiny-scale variant (TinyMLOps)-targeting the resource-constrained microcontrollers embedded in cyber-physical systems (CPS)-remains poorly understood in industrial practice. Opaque models, noisy heterogeneous data, and tight memory budgets hinder adoption in safety-critical settings, where most decisions still rely on human experts. We report a field study of an end-to-end, knowledge-centered TinyMLOps pipeline that fuses domain physics, expert speculation, and sensor streams to deliver explainable, low-footprint models deployable on-device. The pipeline spans automated collection and cleaning of heterogeneous time series, knowledge-driven feature construction, interpretable regularized models, and rolling temporal cross-validation under concept drift. We evaluate it on 4.4 GB of data from two offshore-wind cable-trenching campaigns. The classifier anticipates harmful load peaks up to three minutes ahead at 0.84 AUC within a 32 kB footprint on an ARM Cortex-M4; an ablation shows that injecting prior knowledge halves false alarms and surfaces actionable operational rules. Replaying recommendations in operational dashboards indicates an 11% reduction in non-productive time. We distill engineering lessons and validity threats for trustworthy TinyMLOps in CPS, and release code and an annotated dataset to support reproducibility.

Problem

• TinyMLOps
• Cyber-Physical Systems
• Resource-Constrained Devices
• Explainable AI
• Safety-Critical Systems

Innovation

• TinyMLOps
• knowledge-driven modeling
• explainable AI
• resource-constrained CPS
• concept drift