In this paper, two novel reconfiguration protocols are proposed for an embedded agent, wireless control network for industrial automation systems. This work is motivated by the need for cyber-physical systems for advanced manufacturing that are capable of quickly responding to change while maintaining stable and efficient operation. These protocols are evaluated using a Java Agent DEvelopment (JADE) simulation environment to determine their relative quality of service, bandwidth usage, and reconfiguration efficiency. The results of the experiments support the use of embedded agent technology in low-level devices such as wireless sensor and actor nodes. In particular, device level intelligence provides a direct means to adapt to disturbances on the shop floor while maintaining efficient overall sensing and control system operation.

This paper reports on an ad hoc wireless sensor network architecture for industrial sensing and control applications. This approach is tested using a large-scale (400–676 node) agent-based simulation of a factory environment that is subject to noise and blockage. The basic problem tackled by the distributed system is mobile node tracking. To support this work, we introduce two classes of metrics: (1) a set of tracking performance metrics, and (2) a set of network architecture efficiency metrics. The results of our experiments show that the proposed distributed system adapts readily to changes in the sensing environment, but this higher level of adaptability is at the cost of overall efficiency.

In recent years, there has been a growing interest in flipped delivery of undergraduate courses. There has also been an interest in blending online learning with traditional, in-class learning. In this paper, the efficacy of a blended online course is assessed based on the second-year mechanical engineering course “Computing Tools for Engineering Design” for the Fall 2016 semester. This is an extension of a Fall 2015 study in the same course where traditional lectures were used. This study examines how the online modules are used by the students, as well as students’ opinions on the video effectiveness. The results of the study painted a picture of a typical flipped delivery student: one who streams the content on a personal device/computer before the in-class session, and tends to stop/rewind the content rather than playing it continuously. Student impressions of the mode of delivery were generally positive, indicating that a combination of online lectures and in-class practice sessions support learning.

Advances in cyber-physical systems and the introduction of Industry 4.0 have opened the door for interconnectivity in the industrial automation paradigm. One of the emerging technologies proven to be useful in factory automation is wireless sensor networks. In dynamic situations, wireless sensor networks need to be able to self-reconfigure while maintaining data integrity and efficiency. One solution popular with researchers is the use of multi-agent systems to manage wireless sensor networks. Typically, software agents are located on a server or cloud environment. Recent advances in microcomputers have made it feasible to embed these agents on the devices they control. This requires new reconfiguration and network management protocols. In this paper, an embedded agent architecture for wireless sensor network is proposed and an application specific example is given for an oil and gas refinery. An experiment is also conducted to investigate the effect of cluster sizes and signal frequency on the ratio of lost signals in a wireless sensor network cluster.

In this paper, we describe a distributed clustering technique for wireless sensor node tracking in an industrial environment. The research builds on extant work on wireless sensor node clustering by reporting on: (1) the development of a novel distributed management approach for tracking mobile nodes in an industrial wireless sensor network; and (2) an objective assessment of the cluster management approach both in terms of its tracking performance and its use of network resources. To support this work, we introduce two classes of metrics: a set of three tracking performance metrics, and a set of three network efficiency metrics. The results of our experiments show that the proposed distributed system adapts readily to changes in the sensing environment, but this higher level of adaptability is at the cost of overall efficiency.

With advances in cyber-physical systems and the introduction of industry 4.0, there has been an extensive amount of research in distributed intelligent control. Because in cyber-physical systems it is required that devices are aware of their environment, industrial wireless sensor networks are considered for such types of applications. In this paper, a sink node-embedded multi-agent system is proposed in order to manage clusters of wireless sensors; this architecture is analysed in an oil and gas refinery example.

Wireless sensor networks can be used to monitor many different applications, including factory automation. High-variety manufacturing is becoming the new norm, and requires systems capable of rapidly reconfiguring without downtime. Homogeneous wireless sensor networks may provide a system capable of minimal setup and change-over effort due to their hardware indifferences, but are constrained by limited node lifetime. In this research a distributed, a multi-agent cluster management system that attempts to regulate node death, encouraging maintenance approaches to reclaim and replace nodes is proposed. This system allows the network to continuously operate, and eliminates any downtime due to wireless node power loss. In this paper, an overview of a wireless sensor network management approach, the simulator platform developed using Java Agent Development Environment (JADE) that will be used to test this design, as well as preliminary experimental results using this simulator are proposed. The experimental results focus on the problem of tracking mobile nodes in an industrial wireless sensor network where wireless nodes are subject to battery depletion. The results of preliminary experiments with a heterogeneous network (i.e., a network with pre-defined cluster heads) show that a distributed, multi-agent systems approach is capable of adapting to disturbances resulting from node loss.

Wireless Sensor Networks (WSNs) deployed on a shop floor for factory automation are subject to dynamic and uncertain conditions present with heavy machinery. To overcome these obstacles, on one hand, WSNs should be reconfigurable and adaptable to changes in the shop floor and ambient condition; on the other hand, WSNs should be efficient in consuming the limited wireless network resources. To accomplish these goals, we propose two cluster-based architectures (i.e. static or dynamic clusters), to organize the overall shop floor into a set of tracking zones, each composed of a sink node and a set of closest corresponding anchor nodes. To manage the wireless nodes activities and inter and intra cluster communications, an agent-based technique is employed. To compare the architectures, we report on a set of experiments performed in JADE (Java Agent Development Environment). In these experiments, we compare two agent-based approaches (dynamic and static) for managing clusters of wireless sensor nodes in a distributed tracking system. The experimental results corroborate the efficiency of static clusters versus the robustness of the dynamic clusters.