Industrial assets often feature multiple sensing devices to keep track of their status by monitoring certain physical parameters. These readings can be analyzed with machine learning (ML) tools to identify potential failures through anomaly detection, allowing operators to take appropriate corrective actions. Typically, these analyses are conducted on servers located in data centers or the cloud. However, this approach increases system complexity and is susceptible to failure in cases where connectivity is unavailable. Furthermore, this communication restriction limits the approach’s applicability in extreme industrial environments where operating conditions affect communication and access to the system. This paper proposes and evaluates an end-to-end adaptable and configurable anomaly detection system that uses the Internet of Things (IoT), edge computing, and Tiny-MLOps methodologies in an extreme industrial environment such as submersible pumps. The system runs on an IoT sensing Kit, based on an ESP32 microcontroller and MicroPython firmware, located near the data source. The processing pipeline on the sensing device collects data, trains an anomaly detection model, and alerts an external gateway in the event of an anomaly. The anomaly detection model uses the isolation forest algorithm, which can be trained on the microcontroller in just 1.2 to 6.4 s and detect an anomaly in less than 16 milliseconds with an ensemble of 50 trees and 80 KB of RAM. Additionally, the system employs blockchain technology to provide a transparent and irrefutable repository of anomalies.
Everyday objects are becoming smart enough to directly connect to other nearby and remote objects and systems. These objects increasingly interact with machine learning applications that perform feature extraction and model inference in the cloud. However, this approach poses several challenges due to latency, privacy, and dependency on network connectivity between data producers and consumers. To alleviate these limitations, computation should be moved as much as possible towards the IoT edge, that is on gateways, if not directly on data producers. In this paper, we propose a design framework for smart audio sensors able to record and pre-process raw audio streams, before wirelessly transmitting the computed audio features to a modular IoT gateway. In this paper, an anomaly detection algorithm executed as a micro-service is capable of analyzing the received features, hence detecting audio anomalies in real-time. First, to assess the effectiveness of the proposed solution, we deployed a real smart environment showcase. More in detail, we adopted two different anomaly detection algorithms, namely Elliptic Envelope and Isolation Forest , that were purposely trained and deployed on an affordable IoT gateway to detect anomalous sound events happening in an office environment. Then, we numerically compared both the deployments, in terms of end-to-end latency and gateway CPU load, also deriving some ideal capacity bounds.
Inspection of mining assets is a crucial part of the maintenance process and is of interest to several stakeholders (e.g., OEMs, owners, users, and inspectors). Inspections require an inspector to verify several characteristics of the assets onsite, typically using legacy and poorly digitized procedures. Thus, many research opportunities arise from the adoption of digital technologies to make these procedures more efficient, reliable, and straightforward. In addition to cloud computing, the ubiquitous presence of modern mobile devices, new measurement tools with embedded connectivity capabilities, and blockchain technologies could greatly improve trust and transparency between the stakeholders interested in the inspection. However, there has been little discussion on integrating these technologies into the mining domain. This paper presents and evaluates an end-to-end system to conduct inspections using mobile devices that directly interact with constrained IoT sensor devices. Furthermore, our proposal provides a method to integrate constrained IoT devices as smart measuring tools that directly interact with a blockchain system, guaranteeing data integrity and increasing the trustworthiness of the data. Finally, we highlight the benefits of our proposed architecture by evaluating a real case study in a mining inspection scenario.
Empowering the Internet of Things devices with Artificial Intelligence capabilities can transform all vertical applications domains within the next few years. Current approaches favor hosting Machine Learning (ML) models on Linux-based single-board computers. Nevertheless, these devices’ cost and energy requirements limit the possible application scenarios. Conversely, today’s available 32-bit microcontrollers have much lower costs and only need a few milliwatts to operate, making them an energy-efficient and cost-effective alternative. However, the latter devices, usually referred to as far edge devices, have stringent resource constraints and host non-Linux-based embedded real-time operating systems. Therefore, orchestrating such devices executing portions of ML applications represents a major challenge with current tools and frameworks. This paper formally introduces the Tiny-MLOps framework as the specialization of standard ML orchestration practices, including far edge devices in the loop. To this aim, we will tailor each phase of the classical ML orchestration loop to the reduced resources available onboard typical IoT devices. We will rely on the proposed framework to deliver adaptation and evolving capabilities to resource-constrained IoT sensors mounted on an industrial rotary machine to detect anomalies. As a feasibility study, We will show how to programmatically re-deploy ML-based anomaly detection models to far edge devices. Our preliminary experiments measuring the system performance in terms of deployment, loading, and inference latency of the ML models will corroborate the usefulness of our proposal.