The rise of emerging technologies and Internet of Things (IoT) applications in industrial systems is facilitating the development of Industrial IoT (IIoT). In the industrial sector, IIoT serves as a novel vision of IoT by automating smart objects for processing, communicating, collecting, and sensing real-time events in different industrial systems. This article discusses IIoT and its role in transforming manufacturing processes.
Image Credit: Summit Art Creations/Shutterstock
An Overview of IIoT
IIoT is the network of highly connected and intelligent industrial components deployed to realize high production rates with decreased operational costs through real-time monitoring, efficient controlling and management of industrial assets and processes, operational time, and predictive maintenance.
A subset of IoT, IIoT requires enhanced levels of security and safety and reliable communication without disruptions in real-time industrial operations due to the mission-critical nature of industrial environments. Industry 4.0, a subset of IIoT, specifically focuses on efficiency and safety in manufacturing. The IIoT will evolve extensively in future industrial networks to drive innovation, enabling Industry 5.0 systems to reduce the gap between machines and humans and playing a critical role in realizing the massive personalization vision of Industry 6.0.
Industries can collect and analyze vast amounts of data using IIoT to improve overall industrial system performance. Reduction in capital expenditures is another major benefit of using IIoT. A generic IIoT architecture consists of four layers including layer 1, layer 2, layer 3, and layer 4.
Industrial data sources and IIoT devices generate continuous data streams at layer 1. Edge servers and cloud computing systems/cloud platforms empower IIoT applications at layer 2 and layer 3, respectively. The cloud platform at layer 3 contains information domains and operation domains. Enterprise operations comprising business domains and application domains are performed at layer 4.
Information and data flow among various layers in this architecture, with the operational flow concerning asset management and orchestration flow concerning resource management in the industrial networks. However, IIoT architectures are often perceived differently considering design variations based on resilience, addressability, privacy, security, safety, resource management schemes, execution paradigms, computational assignments, communication paradigms, and location awareness.
Enabling Technologies for IIoT
The IIoT backbone is established by enabling many technologies, including big data analytics, cloud computing, IoT, cyber-physical systems (CPS), artificial intelligence (AI), virtual reality, augmented reality, machine-to-machine (M2M), and human-to-machine (H2M) communication.
In a connected factory scenario, IoT devices enable real-time data collection and actuation. Being the primary IIoT component, these IoT devices track factory assets around the world. Specifically, the entire process starting from raw materials and ending with finished products is monitored using IoT devices to realize a substantial reduction in manual system management and labor cost.
In a fully connected IIoT system, IoT devices are deployed in all factory facilities, including warehouses, distribution centers, and production sites. However, the deployment, maintenance, and monitoring of these devices is a difficult task and necessitates highly qualified technical staff.
AI techniques ensure the autonomous and intelligent functioning of the IIoT system to improve efficiency and minimize human interventions. Complex AI technologies like conversational AI and multi-agent systems are utilized to make the IIoT autonomous.
Additionally, the intelligence is embedded at layers, from sensors and devices to cloud data centers and edge servers, in IIoT systems by enabling various prediction, search, and optimization algorithms. IIoT systems empower diverse CPSs like industrial robots to reduce human interventions and efforts.
The CPS core lies on onboard embedded IoT devices, which allow various actuators and sensors to operate in industrial environments. These onboard embedded IoT devices facilitate intelligent data processing for greater efficiency and autonomous operations in IIoT systems.
Augmented reality technologies assist industrial workers during complex operations like assembling/disassembling the machinery and in tasks in mission-critical systems. These technologies monitor the machines and workers during operations and generate notifications immediately to minimize errors.
Virtual reality technologies facilitate the visualization of configurations and re-configurations of industrial modules and functions before practical implementations in IIoT systems. There-configuration times are reduced and the shut-down time of industrial machines and plants is curtailed using virtual reality.
IIoT Applications
An IIoT suite was proposed to realize the re-industrialization of Hong Kong by mitigating diverse challenges like establishing a network system that enables objects to communicate between the network and other objects in real-time and real-time identification of objects and their locations throughout the manufacturing processes.
High production and upgradation of the manufacturing industry are realizable using this IIoT suite. A cloud platform and a smart hub are the key components of the proposed IIoT suite. The smart hub serves as an IoT device gateway and manages IoT devices at various locations.
This hub accomplishes three tasks, including enabling the data processing, data exchange, and communication between IoT devices; providing convenient solutions during system scaling with new IoT devices; and offering a secure connection channel between the cloud platform and IoT devices by performing data filtering, collection, formatting, and aggregation.
Additionally, the IIoT's cloud platform performs identification and access management, controlling and monitoring IoT devices, routing algorithms, device discovery/configuration, and load balancing. The configuration, deployment, and interaction between heterogeneous IIoT devices are important issues in the manufacturing industry. To address these issues, an IIoT-based hub, designated as IIHub, containing three modules has been proposed in a study.
The first module/customized access module (CA-Module) connects heterogeneous devices called physical manufacturing resources (PMRs) through a set of communication protocols. Similarly, the second module/access hub (A-Hub) serves as a link between CA-Module, smart terminals, and factory workers through constrained application protocol (CoAP) and Wifi/Ethernet interfaces.
The most important third module/smart terminals/local pool service (LPS) performs multiple functions like data processing, storing, and collection, and smart decision-making. Based on the PMR-generated data, LPS processes data in real-time and predicts the total energy consumption, expected production rate, and PMRs predicted maintenance.
Every IIHub module is embedded using special-purpose libraries. For instance, the groups of communication protocols in the CA-Module interact between them using the communication protocol package library. Similarly, A-Hub possesses an embedded multi-dimensional information models library that assists in connectivity, while LPS has an embedded data processing algorithm library, which performs decision-making and data processing and analysis.
Different communication protocols play a crucial role in IIoT systems. For instance, the ZMQ messaging design model represents a flexible and generic machine-to-machine messaging mechanism between machines for command and event notification and data sharing.
Experiment using a case study of quality inspection microwave sensor has shown the effectiveness of the ZMQ technique in dealing with machine presence and discovery, machine connectivity, and messaging, allowing ubiquitous data interaction and access for rich sensing IoT applications. Thus, the proposed method addresses the heterogeneity and complex structure problems of IIoT applications and contributes to cross-platform capability, allowing the implementation on different lightweight devices and computers.
New Developments and Conclusion
A paper published in IEEE Industrial Electronics Magazine explored the feasibility of edge AI, which is a combination of AI with edge computing, for IIoT applications. Privacy preservation, responsiveness, and personalization are the three key edge AI challenges.
Thus, researchers proposed a federated active transfer learning (FATL) model, which can address the edge AI challenges of personalization, responsiveness, and privacy preservation using active learning (AL), transfer learning (TL), and federated learning (FL), respectively.
Specifically, AL personalized the AI model by altering the amount of labeled samples corresponding to task requirements. TL increased responsiveness by quickly adapting the model to new learning tasks. FL ensured privacy due to a distributed training process where no data is shared among devices.
The simulation results demonstrated that the FATL global model could effectively address the edge AI open challenges for future IIoT applications. FATL achieved higher accuracy with a low number of federated devices and maintained high accuracy even when the amount of training samples was reduced significantly. Moreover, the FATL training process required significantly less time compared to other solutions.
Another paper published in IEEE Transactions on Network Science and Engineering investigated the issue of facilitating flexible and efficient federated edge learning (FEEL) in edge-enabled IIoT networks. It proposed a novel FL framework designated as multi-exit-based FEEL (ME-FEEL) to address the system heterogeneity problem in IoT networks. Results from the study showed that the proposed ME-FEEL could effectively outperform the conventional FEEL in resource-limited IoT networks.
To conclude, IIoT is transforming manufacturing through real-time monitoring, increased efficiency, and reduced costs. New developments in AI and federated learning are addressing challenges in personalizing and optimizing IIoT. However, overcoming technical challenges like resource management, reliability, heterogeneity, scalability, interoperability, and security is a must to fully exploit the potential of complex and heterogeneous IIoT systems.
References and Further Reading
Khan, W. Z., Rehman, M. H., Zangoti, H. M., Afzal, M. K., Armi, N., Salah, K. (2020). Industrial internet of things: Recent advances, enabling technologies and open challenges. Computers & Electrical Engineering, 81, 106522. https://doi.org/10.1016/j.compeleceng.2019.106522
Malik, P. K., Sharma, R., Singh, R., Gehlot, A., Satapathy, S. C., Alnumay, W. S., Pelusi, D., Ghosh, U., Nayak, J. (2021). Industrial Internet of Things and its Applications in Industry 4.0: State of The Art. Computer Communications, 166, 125-139. https://doi.org/10.1016/j.comcom.2020.11.016
Foukalas, F., Tziouvaras, A. (2021). Edge artificial intelligence for industrial internet of things applications: An industrial edge intelligence solution. IEEE Industrial Electronics Magazine, 15(2), 28-36. https://doi.org/10.1109/MIE.2020.3026837
Tang, S., Chen, L., He, K., Xia, J., Fan, L., Nallanathan, A. (2022). Computational intelligence and deep learning for next-generation edge-enabled industrial IoT. IEEE Transactions on Network Science and Engineering. https://doi.org/10.1109/TNSE.2022.3180632