The Internet of Things (IoT) is a network of electrical devices that are connected to the Internet wirelessly. This group of devices generates a large amount of data with information about users, which makes the whole system sensitive and prone to malicious attacks eventually. The rapidly growing IoT-connected devices under a centralized ML system could threaten data privacy. The popular centralized machine learning (ML)-assisted approaches are difficult to apply due to their requirement of enormous amounts of data in a central entity. Owing to the growing distribution of data over numerous networks of connected devices, decentralized ML solutions are needed. In this paper, we propose a Federated Learning (FL) method for detecting unwanted intrusions to guarantee the protection of IoT networks. This method ensures privacy and security by federated training of local IoT device data. Local IoT clients share only parameter updates with a central global server, which aggregates them and distributes an improved detection algorithm. After each round of FL training, each of the IoT clients receives an updated model from the global server and trains their local dataset, where IoT devices can keep their own privacy intact while optimizing the overall model. To evaluate the efficiency of the proposed method, we conducted exhaustive experiments on a new dataset named Edge-IIoTset. The performance evaluation demonstrates the reliability and effectiveness of the proposed intrusion detection model by achieving an accuracy (92.49%) close to that offered by the conventional centralized ML models’ accuracy (93.92%) using the FL method.

Most of wireless sensor networks require time synchronization to wake up sensor nodes at scheduled time. It helps accomplish coordinated communications among the nodes to save energy. Wireless sensor nodes are often equipped with a low-cost clock oscillator with large clock drift. Thus, it is important to correct the long-term synchronization error caused by clock drift. This paper introduces a low power protocol for long-term synchronization called Fast Scheduling and Accurate Drift Compensation for Time Synchronization (FADS). It chooses a minimal sequence of synchronization steps by estimating the density of the neighbor nodes. FADS achieves low power synchronization by reducing the number of message exchanges and increasing the synchronization interval (the period in which each node needs to perform synchronization). To evaluate the performance of FADS, we implemented it in real wireless networks with sensor hardware and conducted extensive tests. This paper demonstrates that FADS achieves substantially fewer messages and less delays in the scheduling phase than a previous technique, Density Table based Synchronization (DTSync). Simulation results show that FADS reduces the number of messages by 50 times and reduces the delay by 4 times compared to DTSync for large networks. In addition, experiments with actual sensor devices demonstrate that it allows significantly less frequent synchronization 195 times improvement compared to DTSync.

The growth in the Industrial Internet of Things (IIoT) drives the development of large-scale wireless sensor networks (WSN), such as smart cities. Time synchronization is essential and primary for enabling collaborative operations among sensor nodes. Generally, all sensor nodes in the network synchronize through exchanging messages, increasing energy consumption. This study introduces scheduling-based low-energy synchronization (SLES) for IIoT. SLES aims to reduce the number of transmitted messages, which in turn reduces the overall energy consumption. In addition, SLES establishes a schedule of distinct synchronization messaging times for each reference node to alleviate the collision problem among the sensor nodes. SLES is constructed and evaluated using real WSNs with different configurations. The experimental results showed that SLES achieves significantly less message traffic than the previous algorithms: FADS – fast scheduling and accurate drift compensation for time synchronization, EERS – energy-efficient reference node selection, and R-sync – robust time synchronization. In this study, we used a simulation to evaluate the performance of SLES in large-scale networks. For the 4-way grid network, the experiential results show that SLES reduced the number of transmitted messages by approximately 22% compared to EERS and FADS. Moreover, SLES outperformed the R-sync algorithm by a factor of 2.5.

Time synchronization is becoming a general requirement for wireless sensor networks to wake up sensor nodes and achieve successful communications among the nodes. It is, however, challenging for multi-hop sensor networks to synchronize all the distributed nodes autonomously with minimum power consumption. This paper proposes a low power time synchronization protocol called Density Table based Synchronization (DTSync). It selects a minimal sequence of synchronization steps by estimating the density of the neighboring nodes and mapping these nodes in a density table. The proposed protocol attempts to minimize the power consumption by reducing the number of messages exchanged among the nodes. We evaluate the proposed protocol using software simulation and then real wireless networks with sensor hardware. Compared to a prior method such as Hierarchy Reference Broadcast Synchronization (HRTS), DTSync achieves substantially fewer synchronization messages and thus lower power consumption. In addition, DTSync is guaranteed to synchronize all nodes in any multi-hop networks, whereas the previous methods are not guaranteed to synchronize all nodes. Simulation experiments demonstrate that DTSync reduces the number of synchronization messages by more than 40 times compared to HRTS for a large network of 1500 node.