Networks deployed in real-world conditions have to cope with dynamic, unpredictable environmental temperature changes. These changes affect the clock rate on network nodes, and can cause faster clock de-synchronization compared to situations where devices are operating under stable temperature conditions. Wireless network protocols, such as time-slotted channel hopping (TSCH) from the IEEE 802.15.4-2015 standard, are affected by this problem, since they require tight clock synchronization among all nodes for the network to remain operational. This paper proposes a method for autonomously compensating temperature-dependent clock rate changes. After a calibration stage, nodes continuously perform temperature measurements to compensate for clock drifts at runtime. The method is implemented on low-power Internet of Things (IoT) nodes and evaluated through experiments in a temperature chamber, indoor and outdoor environments, as well as with numerical simulations. The results show that applying the method reduces the maximum synchronization error more than ten times. In this way, the method allows reduction in the total energy spent for time synchronization, which is practically relevant concern for low data rate, low energy budget TSCH networks, especially those exposed to environments with changing temperature. 

The digital twin is a rather new industrial control and automation systems concept. While the approach so far has gained interest mainly due to capabilities to make advanced simulations and optimizations, recently the possibilities for enhanced security have got attention within the research community. In this article, we discuss how a digital twin replication model and corresponding security architecture can be used to allow data sharing and control of security-critical processes. We identify design-driving security requirements for digital twin based data sharing and control. We show that the proposed state synchronization design meets the expected digital twin synchronization requirements and give a high-level design and evaluation of other security components of the architecture. We also make performance evaluations of a proof of concept for protected software upgrade using the proposed digital twin design. Our new security framework provides a foundation for future research work in this promising new area.

Federated Learning (FL) is a promising paradigm to empower on-device intelligence in Industrial Internet of Things(IIoT) due to its capability of training machine learning models across multiple IIoT devices, while preserving the privacy of their local data. However, the distributed architecture of FL relies on aggregating the parameter list from the remote devices, which poses potential security risks caused by malicious devices. In this paper, we propose a flexible and robust aggregation rule, called Auto-weighted Geometric Median (AutoGM), and analyze the robustness against outliers in the inputs. To obtain the value of AutoGM, we design an algorithm based on alternating optimization strategy. Using AutoGM as aggregation rule, we propose two robust FL solutions, AutoGM_FL and AutoGM_PFL. AutoGM_FL learns a shared global model using the standard FL paradigm, and AutoGM_PFL learns a personalized model for each device. We conduct extensive experiments on the FEMNIST and Bosch IIoT datasets. The experimental results show that our solutions are robust against both model poisoning and data poisoning attacks. In particular, our solutions sustain high performance even when 30% of the nodes perform model or 50% of the nodes perform data poisoning attacks.

Cellular networks are envisioned to be a cornerstone in future industrial Internet of Things (IIoT) wireless connectivity in terms of fulfilling the industrial-grade coverage, capacity, robustness, and timeliness requirements. This vision has led to the design of vertical-centric service-based architecture of 5G radio access and core networks. The design incorporates the capabilities to include 5G-AI-Edge ecosystem for computing, intelligence, and flexible deployment and integration options (e.g., centralized and distributed, physical, and virtual) while eliminating the privacy/security concerns of mission-critical systems. In this article, driven by the industrial interest in enabling large-scale wireless IIoT deployments for operational agility, flexible, and cost-efficient production, we present the state-of-the-art 5G architecture, transformative technologies, and recent design trends, which we also selectively supplemented with new results. We also identify several research challenges in these promising design trends that beyond-5G systems must overcome to support rapidly unfolding transition in creating value-centric industrial wireless networks.

This paper presents the use of automated machine learning for solving a practical problem of a real-life Smart Waste Management system. In particular, the focus of the paper is on the problem of detection (i.e., binary classification) of emptying of a recycling container using sensor measurements. Numerous data-driven methods for solving the problem are investigated in a realistic setting where most of the events are not actual emptying. The investigated methods include the existing manually engineered model and its modification as well as conventional machines learning algorithms. The use of machine learning allows improving the classification accuracy and recall of the existing manually engineered model from 86.8% and 47.9% to 99.1% and 98.2%, respectively, when using the best performing solution. This solution uses a Random Forest classifier on a set of features based on the filling level at different given time spans. Finally, compared to the baseline existing manually engineered model, the best performing solution also improves the quality of forecasts for emptying time of recycling containers.