The growing integration of sensors and embedded devices into distributed Internet of Things (IoT) systems has increased the demand for real-time data collection and processing solutions. However, the high volume and frequency of sensor data create challenges related to storage, transmission, and response latency, especially in resource-constrained environments. In this context, locally executed compression techniques, aligned with the Tiny Machine Learning (TinyML) paradigm, become differentiators for enabling embedded applications. Thus, this work proposes an autoencoder-based approach for efficiently compressing sensor data on edge devices. Three autoencoder variants (feed-forward, sparse, and contractive) are evaluated, combined with symmetric and asymmetric architectures, considering criteria such as compression ratio, information preservation, and embedded execution feasibility. For practical validation, a case study was conducted using vehicular data collected via the OBD-II interface, where the selected models were deployed on the OBDII Edge Freematics One+ device. The results show that the models could reduce data dimensionality with minimal information loss, maintain competitive performance on discriminative tasks, and exhibit inference times compatible with real-time applications. Autoencoders represent a viable neural compression solution for IoT environments, potentially applicable to various embedded sensing scenarios.

The advancement of Industry 4.0 has intensified the demand for intelligent systems that can efficiently access and interpret technical information critical to industrial operations. However, recovering knowledge from extensive and complex technical documentation remains a significant challenge. This study examines the effectiveness of various Retrieval-Augmented Generation (RAG) strategies, combined with different Large Language Models (LLMs), in extracting and generating answers from industrial technical documents. A case study was conducted based on PROFIBUS, a widely adopted digital communication protocol in automation networks, with technical documents organized into categories for engineers and developers. Twenty questions of varying complexity were formulated, and responses were generated using three RAG strategies (Basic, Decomposition, and HyDE) combined with two LLMs (Gemma 3 and GPT-4o-mini). The outputs were compared against reference answers generated by the NotebookLM system and evaluated using automatic metrics, including ROUGE, METEOR, BERTScore, and MATTR. The results indicate that the Decomposition and HyDE strategies achieved superior semantic similarity scores when combined with more capable models. That model’s performance varied depending on the complexity of the document profile. These findings highlight the importance of tailored RAG strategies in enhancing intelligent information retrieval in industrial domains, which supports safer and more efficient operational environments.

The Internet of Things (IoT) generates massive multivariate time series data requiring real-time anomaly detection and correction for reliable monitoring. This challenges resource-constrained embedded devices due to conventional offline training and batch processing. To address this, we propose two algorithms derived from the TEDARLS framework: Multivariate Sequential TEDA (MST) and Multivariate Parallel TEDA (MPT). Derived from the TEDARLS framework, both enable on-device detection and correction of multivariate anomalies within TinyML constraints. A case study with real vehicular sensor data demonstrated low inference times and consistent embedded behavior. Supervised metrics were only assessed on synthetic data. MPT, though more sensitive, introduced greater signal distortions and required significantly longer processing times. Overall, MST demonstrated superior stability and suitability for real-time anomaly correction in resource-constrained IoT environments. This approach addresses an important gap in embedded analytics for IoT by enabling lightweight, accurate, and autonomous anomaly detection and correction at the edge.

Respecting real-time requirements in networked control systems requires that the network traffic respects some well-defined constraints. Any abnormal behavior, e.g. due to malfunctioning of few devices or representing a malicious attack, could disrupt the overall plant behavior. Immediate recognition of anomalies should occur, to determine the causes and mitigate the impairments. Despite hardware sniffer being ready available, deployment cost, maintenance cost and topology constrains of industrial network discourage their adoption for live traffic collection. Additionally, anomaly detection systems are typically complex solutions that demand experienced personnel, which is not generally available among plant maintenance workers. This paper presents the ongoing research activity for a simple but effective expert system that gathers network-related figure of merits natively evaluated by industrial-grade devices, acquired by standard SNMP (Simple Network Management Protocol) queries. Data collected by such a distributed measurement system are further analyzed using simple classification techniques, as the unsupervised clustering, which limit the computational burden. In this work some suitable performance indicators are addressed and the related multi dimensional clustering is discussed. Some preliminary results are provided, exploiting a reference use case designed around a real-world assembly machine.

The Industry 4.0 has led to a significant transformation in manufacturing and industrial processes, characterized by increased connectivity and widespread digitalization. This revolution is driven by the convergence of IT (Information Technology) paradigms with OT (Operational Technology) domains. Virtualization technology plays a crucial role in this integration, thanks to its characteristics of agility, scalability, and efficiency. This work investigates the virtualization of Programmable Logic Controllers (PLCs) and it is focused on the need for evaluating their performance. PLCs are essential components of industrial automation and are being virtualized to increase flexibility, reduce hardware dependencies, and enable seamless integration with IT infrastructures. This work proposes a methodology for evaluating the execution time of a benchmark program and the PLC real-time behavior. Experiments on a real use case demonstrate that the PLCs implemented in containers may be 50 times faster than the real ones in executing data elaboration, while no significant differences were observed in real-time communication. In any case, virtual PLCs are characterized by greater jitter with respect to real PLCs.

The demand for enhanced flexibility in automation systems, dictated by initiatives as the Industry 4.0 in Europe or the Industrial Internet Consortium in the United States, can be satisfied inheriting solutions of the information technology (the so called IT-OT convergence). Other than communication and cloud technologies, virtualization will play a relevant role in the next future. The use of lightweight virtualization, e.g., in the form of containers for implementing the edge computing paradigm, is a promising tool to dynamically provide services depending on the actual production needs. However, requirements of industrial automation applications are different from those of office-like counterparts. For this reason, methodologies for evaluating obtainable performance are needed. In this work, an industrial-grade framework, the Siemens Industrial Edge, is considered. A reference testbed is setup to evaluate latencies when information is exchanged between two user applications. Experiments demonstrate that round trip time in the order of 10 ms is feasible, compatible with the requirements of both process control systems and supervision applications.