Digital transformation and visual knowledge map analysis of intelligent factory for sensor information of internet of things

Abstract

Industry 4.0 is reshaping conventional factories into “smart factories” via the widespread use of IoT-enabled networks of linked devices, sensors, and software for process optimization and monitoring. Intelligent manufacturing facilities may employ IoT-based predictive maintenance to reduce downtime, increase equipment longevity, and avoid machine problems. Manufacturers may get real-time insights into energy consumption patterns, which is a major concern in the business. The primary objective is to optimize energy use during part manufacturing. Hence, this paper proposes the Internet of Things- Low-Power Wide-Area Network Model (IoT-LPWM) to monitor manufacturing and reduce energy consumption. The proposed method's production status component uses visual Knowledge Map Analysis loaded with data from the edge device. A Low-power wide-area network (LPWAN) is the fundamental component of the suggested approach to industrial wireless communication. Using edge computing technology in LPWAN helps reduce computational complexity by shifting high-intensity processing to the periphery, where devices with computing resources are more readily available. Both the energy needed to process and store massive data and the likelihood of cyberattacks may be decreased with this method. The experimental results show that the IoT-LPWM provides useful information to help them make decisions and reduce energy consumption. The experimental results show that our proposed method IoT-LPWM achieves a high performance ratio of 97%, attack prevention ratio of 96.3%, energy management ratio of 93.8%, and data transmission ratio of 98.1% compared to other methods.

Keywords

Intelligent factory internet of things knowledge map edge computing sensors

1. Significance of industrial internet of things (IIoT) for smart manufacturing

Intelligent factories are an integral feature of Industry 4.0, the fourth industrial revolution, which aims to improve manufacturing efficiency.¹ Internet of Things (IoT), artificial intelligence (AI), and robots are some of the state-of-the-art technologies that are being integrated into industrial processes as part of Industry 4.0 to boost efficiency and adaptability.² A “smart factory” in the context of Industry 4.0 is a completely automated manufacturing facility that utilizes state-of-the-art technologies such as robotics, AI, and the IoT to improve quality, efficiency, and productivity.³ A “smart factory” is a manufacturing facility with integrated digital technology to maximize output while decreasing waste.⁴ Smart factories can streamline production processes and increase operational efficiency by utilizing cutting-edge technology like the IoT, artificial intelligence (AI), and robots.⁵ An Industry 4.0 smart factory aims to facilitate real-time monitoring, analysis, and decision-making via networked and communicative machinery and equipment.⁶ In this manner, factories may streamline their operations, cut down on emissions, and be more adaptable to meet the needs of a dynamic market.⁷ To facilitate digital transformation in manufacturing and help businesses run more efficiently, cheaply, and sustainably, the Internet of Things (IoT) has lately been integrated into smart factories.⁸ Smart factories are becoming more commonplace because of the Internet of Things. In addition, the data is created from other sources, which are then combined and layered.⁹ Conventional big data analytics approaches have several drawbacks when dealing with large and complicated data sets, including excessive energy consumption, inconsistent latency, insufficient data fusion, and inadequate security.¹⁰ The Internet of Things (IoT) and edge computing are crucial enabling technologies in the modern industrial environment since sensors are the primary data source.¹¹ To improve data transfer and the performance of applications that need real-time processing, edge computing enables computation to be executed near physical devices.¹²

The business world necessitates a data connection system with enough real-time processing capacity, reliability, stability, and fault tolerance to access the hidden information therein.¹³ Edge devices have become more potent concerning processing speed, memory capacity, and the number of embedded functionality. As a result, it now offers safe, low-energy, and latency-sensitive application support by moving data, application, and processing power closer to the user.¹⁴ The capacity to learn and find new things based on big data and each other is present at every edge of this IoT ecosystem.¹⁵

Processing and analyzing enormous applications at the edge allows for the computation of difficult tasks in almost no time. Artificial intelligence (AI) applications may be trained and deployed at the edge because of the massive data collection. An artificial intelligence (AI) application for production planning processes data sent by the machine's Numerical Control Unit (NCU) to a cloud platform. Moving away from preventative maintenance and toward predictive techniques has been made possible by the massive amounts of data lately available. Visual Knowledge map analysis allows edge intelligence called predictive analytics and production statistics created from subsets of the acquired data rather than the complete datasets as needed and desired. Visual knowledge maps are increasingly used in business to help with data collection, allowing quicker and more accurate development of AI systems. Visual Knowledge maps are very useful in building engineering for making incremental improvements. A system for tracking a machine's actions has been implemented using a set of concepts.

Industrial environments, including manufacturing lines, robots, and CNC machines, provide smart sensors with the raw data they need to gather. Smart sensors combine and pre-process data in several dimensions. Smart gateways then take the pre-processed data and use it to ascertain the knowledge of the smart sensors and industrial environment. At last, the data undergoes processing before being sent to the cloud. This ensures that only the crucial processed information is posted. The architecture is an information system because of the Visual Knowledge map chains that run through it. According to this method, the computation and data-communication capabilities of LPWAN and edge devices have been fully realized. Consequently, the intelligent factory might have more explicable results, and the monitored parts of a cyber-physical production system could gain more knowledge from a mix of semantic technologies and data-driven AI methodologies.

The novelty of the paper:

The study presents the IoT-LPWM as a unique industrial monitoring and energy-saving method. This method increases industrial wireless communication using edge computing and LPWAN. Edge device data makes Visual Knowledge Map Analysis stand out for providing decision-makers with full visual insights into production. The approach simplifies processing by moving computationally demanding jobs to the edge, where computer resources are more abundant. This technology saves energy during production while processing and storing large data sets. Decentralized processing reduces cyberattacks, improving cybersecurity. Experimental results showed that the IoT-LPWM model may be used for energy-efficient factory monitoring in industrial IoT. The model reduces energy use and offers crucial decision-making data.

The main contribution of this paper:

i)
Designing the proposed IoT-LPWM to monitor the manufacturing industry and reduce energy consumption.
ii)
The proposed method utilizes visual knowledge map analysis and edge devices to provide real-time production status and store the data in the cloud platform.
iii)
The numerical results and the proposed method IoT-LPWM have been performed to achieve a high-performance ratio compared to other methods.
The upcoming section is as follows: section 2 deliberates the related works, section 3 examines the proposed methodology, section 4 discusses the numerical analysis and Section 5 concludes the overall paperwork.
2. Related works

Innovative integrated method (IIM) to the q-rung ortho-pair fuzzy sets (q-ROFSs) multi-objective optimization based on ratio analysis plus completes multiplicative form (MULTIMOORA) and criterion interaction via inter-criteria correlation (CRITIC).¹⁶ The suggested approach uses CRITIC to determine the weights of the attributes and MULTIMOORA to approximate the order of the alternatives on the q-ROFSs. Sensitivity and comparative studies are performed using the suggested method to demonstrate the created framework's capabilities in intelligent IIoT system prioritizing.

Non-orthogonal multiple Access (NOMA) and Mobile Edge Computing (MEC) have shown promise in handling delay-sensitive jobs inside the IIoT network.¹⁷ Improving the task satisfaction ratio is the goal of creating a distributed DRL system that optimizes the choice of task offloading and sub-channel assignments to support the task offloading policy for binary computing. Based on the results of the simulations, the proposed prediction-based-DRL (P-DRL) method may outperform the current systems in terms of the task-satisfaction ratio.

The Efficient Heterogeneous Signcryption Scheme (EHSS) is a computing-based edge solution for the Industrial Internet of Things (IIoT) that allows for certificateless data transmission and multi-message/multi-receiver communication between edge servers and data consumers.¹⁸ To provide confidentiality and immutability, our proposed technique satisfies the computational Diffie-Hellman criterion and the discrete logarithmic assumption for the randomised oracle model. The testing confirmed our theory, showing that our system is better than the current ones.

The decentralized partially observable Markov decision process (Dec-POMDP) and Dual dual-attention assisted deep reinforcement learning (DADR) algorithm is a novel energy-efficient resource allocation algorithm and it utilizes the centralized training and distributed executed (CTDE) framework to be able to provide constrained by resources nodes with the ability to make intelligent decisions.¹⁹ Based on the simulation findings, the suggested DADR method achieves better network lifespan, transmission reliability, and stability than the current resource allocation algorithms.

Attribute-based searchable encrypted scheme (ASES) technique that combines online/offline encryption with outsourcing decryption, which uses a reusable ciphertext pool to lessen processing during encryption and makes use of an edge server to lower the cost for users with limited resources by outsourcing decryption.²⁰ An additional layer of protection for the solution may be implemented by having the designated server verify the cloud server in the accessible area. Lastly, performance and safety research has shown that our system is efficient and satisfies CCA2-secure.

A neutrosophic weighted product method (NWPM) based machine learning model (MLM) was used to determine the trustworthiness of IIoT objects.²¹ The IIoT devices’ gathered geographical knowledge, temporal experience, and behavioural patterns and the resulting model determines the devices’ reliability. The model suggests neutropolic support vector machines (SVM) and neutrosophic K-NN clustering to categorize the retrieved attributes. The suggested neutrosophic SVM algorithm generates the final trust score and accurately identifies the trust boundaries.

The Federated Learning Method (FLM) aims to safeguard IoT networks from malicious assaults.²² Every time FL training is finished, the global server updates the model and sends it to all IoT clients, who then train their local dataset. Each IoT device may preserve its privacy while the whole network improves. They conducted comprehensive testing using a novel dataset named Edge-IIoTset to determine the efficacy of the proposed method.

Efficient lightweight secure authentication scheme (ELSAS) for the viewpoint of IIoT with a focus on people.²³ The proposed method presupposes a registration centre and proposes an automated generation of a node's public and secret keys upon the first network connection. This method may reduce processing overhead and exponential computations while fixing possible vulnerabilities.

Combining the IIoTs with the IoT in the public sector has been a huge boon to the growth and development of Industry 4.0. Blockchain and Federated Learning for Intrusion Detection in Industrial Internet of Things Networks.²⁴ The FL-IDS strategies address the potential for attacks on IIoT systems and provide suggestions for securing them, showcasing the variety and complexity of approaches to handle security risks in IoT and IIoT settings.

Cognitive computing integrates ideas, methodologies, and technology from several fields. IoT applications allow physical and digital items to communicate.²⁵ Smart and secure IoT systems should be developed for dependability, resistance to threats, operational efficiency, energy efficiency, and resource usage. Cognitive computing embedded data solutions let specialists make decisions by processing and analyzing enormous quantities of data from secure IoT devices.

3. Proposed methodology

The IoT revolutionised manufacturing in the fourth industrial revolution, turning traditional factories into “smart factories” that optimize and monitor processes via interconnected devices, sensors, and software networks. A big issue for manufacturers is energy usage patterns; they may gain real-time insights into these trends. The main goal is maximizing energy consumption efficiency while making the parts. To keep tabs on production while cutting down on power use, this study put forth the IoT-LPWM model. Compiled using data from the edge device, the visual Knowledge Map Analysis is used by the production status component of the proposed technique. The proposed method for wireless communication in industrial settings relies on an LPWAN. By moving the processing of high-intensity tasks to the periphery, where devices with computing capacity are more accessible, edge computing technology in LPWAN helps to decrease computational complexity. This approach can reduce the energy requirements for managing and storing large amounts of data and the probability of cyberattacks. The experimental findings demonstrate that the IoT-LPWM offers valuable insights to assist decision-making and decrease energy use.

A physical segment low-power wide area network (LPWAN) part and a cloud section comprise the three primary components of the IIoT architecture shown in Figure 1. The key data sources in the physical segment are the many pieces of industrial machinery, such as computer numerical control (CNC) machines, industrial robots, and other industrial machines. The smart sensors of the LPWM division collect raw data from these several sources. The physical sector also includes crucial information systems.

Figure 1.

Proposed IoT-LPWM.

Cloud and IoT-LPWM sections on the ground unearth information and knowledge, which is then delivered by these platforms. This structure is a component of the LPWAN. To compile raw data, the smart sensors communicate with the many pieces of industrial gear. The data collected by smart sensors that rely on embedded processors undergoes a preprocessing stage. The physical part receives low-level knowledge, such as system condition warning knowledge and the smart gateway receives the pre-processed data. Helping individuals make better decisions is what this is all about. Next, the pre-processed data is sent to the smart gateways for advanced data analytics tests, such as clustering, statistical regression, and classification. During processing, the data is more abstracted. The smart gates generate two separate types of information. An item of data about smart sensors allows for improving the software and hardware parameter settings of smart sensors.

On the other hand, the physical portion of the whole production system will include additional information. Data that has been processed is uploaded into the cloud under the framework established for IIoT. Integrated modelling and synthesis, collaborative diagnostics, and decision-making are all examples of high-level data analysis that may be carried out in the cloud. The Internet of Learning-based low-power wide area network (LPWM) may be improved using cloud-discovered information. The proposed IoT-LPWM, which includes intelligent sensors and gateways, is equipped with the capacity to learn from one another and the cloud. A further benefit of cloud computing is the discovery of artificial intelligence knowledge, which includes self-configuration, self-adjustment, and self-optimization.

The knowledge map analysis is found and flows across smart sensors and gateways in a suggested LPWM network, considering each edge's computational capabilities. Cloud computing and many knowledge chains are part of the proposed system. What follows is an examination of these knowledge networks’ finer points.

The following is a representation of the smart sensors that gather raw data from the physical section (1):

C = [C_{1}, C_{2}, \dots, C_{O}]

(1)

Where C is the whole raw dataset, including all characteristics from

C_{1}, C_{2}, \dots, C_{O} .

Knowledge I, III, and V are the three categories of information about physical labour uncovered by smart sensors, gateways, and the cloud to manage energy management. The based representation for Knowledge I is shown as (2):

\begin{aligned} F_{1}^{l q} & = H_{l q} (C, e_{1}^{l q}) \\ F_{2}^{l q} & = H_{l q} (C, e_{2}^{l q}) \\ F_{m}^{l q} & = H_{l q} (C, e_{m}^{l q}) \end{aligned}

(2)

The function for finding new information about the physical world is denoted as

H_{l q} (.)

and the edge gained knowledge for this world is denoted as

F_{1}^{l q} \dots F_{m}^{l q}

. The number of edge devices is denoted by m. The parameters of the function

H_{l q} (.)

are denoted as

e_{1}^{l q} \dots e_{m}^{l q}

. In addition, smart sensors, which are characterized as (3):

\begin{aligned} F_{1}^{l e} & = H_{l e} (C_{1}^{F}, e_{1}^{l e}) \\ F_{2}^{l e} & = H_{l e} (C_{2}^{F}, e_{2}^{l e}) \\ F_{m}^{l e} & = H_{l e} (C_{m}^{F}, e_{m}^{l e}) \end{aligned}

(3)

Where

F_{1}^{l e} \dots F_{m}^{l e}

are the smart gateway pre-processing outcomes in the processes above and

H_{l e} (.)

is the smart gateway knowledge discovery function used to analyze the attack levels. The parameters of the function

H_{l e} (.)

are denoted as

e_{1}^{l e}

through

e_{m}^{l e}

\begin{aligned} F_{1}^{l q} & = H_{l q} (C_{1}^{F}, F_{1}^{l q}) \\ F_{2}^{l q} & = H_{l q} (C_{2}^{F}, F_{2}^{l q}) \\ F_{m}^{l q} & = H_{l q} (C_{m}^{F}, F_{m}^{l q}) \end{aligned}

(4)

The smart gateway part of the article uncovers two kinds of information. Consider the following knowledge chain representations: one for the physical world (Knowledge III):

\begin{aligned} F_{1}^{l q} & = e_{l q} (F_{1}^{F q}, E_{1}^{l q}) \\ F_{2}^{l q} & = e_{l q} (F_{1}^{F q}, E_{2}^{l q}) \\ F_{m}^{l q} & = e_{l q} (F_{m}^{F q}, E_{m}^{l q}) \end{aligned}

(5)

In this collection of operations,

F_{1}^{l q} \dots F_{m}^{l q}

represent the physical world's knowledge (Knowledge III), and

e_{l q} (.)

is the operation that performs this physical world's information. The parameters of the function

e_{l q} (.)

are denoted as

E_{1}^{l q}

through

E_{m}^{l q}

. Knowledge II refers to smart sensor-specific information and its associated knowledge chains are (6):

\begin{aligned} F_{1}^{e f} & = e_{l q} (F_{1}^{E} L, F_{1}^{E f}) F_{1}^{l f} \\ F_{2}^{e f} & = e_{l q} (F_{2}^{E} L, F_{2}^{E f}) F_{2}^{l f} \\ F_{m}^{e f} & = e_{l q} (F_{m}^{F q}, E_{m}^{l q}) F_{m}^{l f} \end{aligned}

(6)

The knowledge of smart sensors is represented by the

F_{1}^{e f} \dots F_{m}^{e f}

, which is part of Knowledge II. The function for finding new information is

e_{l q} (.)

. The variables of the function

e_{l q} (.)

are

F_{1}^{l f} \dots F_{m}^{l f}

, and they are associated with the inputs of the smart sensor f. Smart sensors gather data, which is then processed again in the smart gateways before being sent to the cloud (7):

\begin{aligned} F_{1}^{l d} & = E_{l d} (F_{1}^{e l}, E_{1}^{l d}) \\ F_{2}^{l d} & = E_{l d} (F_{2}^{e l}, E_{2}^{l d}) \\ F_{n}^{l d} & = E_{l d} (F_{n}^{e l}, E_{n}^{l d}) \end{aligned}

(7)

Where

E_{l d} (.)

is the function that processes data for the cloud and

F_{1}^{l d} \dots F_{m}^{l d}

are the processed data. The total quantity of smart gateways is represented by n. The parameters of the function

E_{l d} (.)

are denoted as

F_{1}^{e l} \dots F_{1}^{e l} .

The main information (Knowledge V) $D^{l q}$ on the cloud pertains to the physical world and is accessed by the function $g_{l q} (.)$ (8):

D^{l q} = g_{l q} (E^{l d}, Δ^{l q})

(8)

Processed data acquired from the smart gateway constitutes the inputs of

g_{l q} (.)

.Knowledge IV which is based on high-level information and may be sent via the cloud to smart gateways is depicted as (9):

\begin{aligned} D_{1}^{l e} & = g_{l e} (E_{1}^{l d}, Δ^{l e}) \\ D_{2}^{l e} & = g_{l e} (E_{2}^{l d}, Δ^{l e}) \\ D_{n}^{l e} & = g_{l e} (E_{n}^{l d}, Δ^{l e}) \end{aligned}

(9)

Where variables

E_{n}^{l d}

and the knowledge discovery function

g_{l e} (.)

are represented by the knowledge IV and

D_{1}^{l e} \dots D_{n}^{l e}

are the smart gateways’ knowledge inputs. Both statements include the bias, denoted as

Δ^{l e}

and

Δ^{l q}

. Based on the information provided about the framework, five different kinds of knowledge map analyses are found in the suggested framework.

Figure 2 illustrates the required knowledge graphs based on the IIoT network security system. The presented framework is compatible with energy efficiency regarding structure and requirements. Data technology, analytics technology, and operation technology are the three main components of the LPWM architecture. Levels one through four of the five functional categories are sensor connection, cyber, cognitive, and configuration. The smart sensors in the proposed system have defined their roles in sensor connection and data-to-information conversion. Here, edge devices’ data-gathering and processing skills shine. Fast data transmission rate, low latency, high dependability, preprocessing capability, and visualization capabilities are anticipated attributes of data and analytical technologies. The suggested smart gates provide cyber-level services using analytical technologies.

Figure 2.

Knowledge map analysis based on IIoT data security.

In addition, the cloud enables basic functions at the cognition and configuration levels—the intelligent and self-configuring manufacturing system. Operation technology is used for intelligent optimization and corporate management to uncover hidden patterns via knowledge map analysis. These graphs pertain to the security of the network. When it comes to maintaining the network security of the IIoT, it is beneficial to determine the extent of the damage caused by network attacks and install and concentrate on preventing potentially damaging network assaults in advance. However, it is important to consider which factors might be included in addition to the kinds of assaults to estimate the degree of injury more precisely. Various types of attacks cause varying degrees of damage.

This application aims to provide an easy-to-understand evaluation of the production status and details on the state of the industry's important equipment pieces. The component's general design is shown in Figure 3.

Figure 3.

Industrial production Status assessment.

The program uses a MQTT broker to get information from the online platform. Once the knowledge graph is generated, the context builder module stores and semantically correlates the data. Concerning the specifics, the experts’ established rules govern how the context builder generates nodes and their connections. At first, the users had to manually enter the historical data into the Knowledge map analysis (KMA). Nonetheless, the context builder dynamically stores the real-time data and the outcomes from the condition assessment and maintenance forecasting component in the KMA.

Furthermore, the Cypher programming language implements the connections among the features. In the next step, the evaluation module checks the KM data and gives the user feedback on the current production state. The program has always relied on a client-server model. They built the front end using the React JavaScript toolkit for building user interfaces. With Flask and Python, the backend side was built. Web platforms and REST API communication protocols were used to establish data connections between clients and servers.

Figure 4.

IoT-LPWM-based energy efficiency management.

An assertion that the Internet of Things (IoT) would make businesses more efficient is supported by some preliminary data. The concept that machinery, automobiles, and other real-world components may be digitized is potent. At this early point, the Industrial Internet of Things is beginning to have a major influence by altering how it is manufactured and delivered, as well as h and The IIoT can start functioning at its full potential, provided legislative steps are taken to promote interoperability, guarantee security, and safeguard privacy and property rights. This is particularly true if leaders fully embrace data-driven decision-making. Systems that are really in existence and are referred to as WSN feature communication capabilities, information and control processing, and monitoring of sensors.

Creating a network that connects several sensors placed with the central area node and transmits information to process it is the second component of the Industrial Internet of Things (IIoT).

Additionally, this central node is responsible for controlling the various settings of sensor devices and ensuring that connectivity with the cloud is maintained. The mode of operation functions as a gateway, or more specifically, a communication bridge between various sensor devices and the cloud-based system.

The cloud, the central system, and the control room comprise the third segment according to setup resources or offer any data. The cloud is an online infrastructure often situated on the Internet. It enables universal access to a shared point that can be used for any purpose. After that, the central processing system stores all of the information in a database and processes it using big data methods, yet another technology has emerged due to the paradigm shift toward smart societies and industries. All pertinent information is reviewed from the central control room, and as a result, suitable judgments are made.

To obtain a true self-power behaviour that is close to the eternal operation, the design and operation of Very Low Power are performed. The capacity to scale and handle large networks enables monitoring of different construction situations. The cost of the nodes is quite cheap, which makes it possible to install them at a reasonable price even for large-scale structures. • Interfaces that are standardized to enable coexistence and cooperation with any systems that are already in place.

IoT-LPWM integrates LPWAN with edge computing for industrial monitoring and energy savings, making it unique. It moves high-intensity processing to edge devices, simplifying and saving energy. Unlike other systems, the model uses Visual Knowledge Map Analysis to improve production insights. By reducing cyberattack risk, its decentralized approach improves cybersecurity. This makes the IoT-LPWM approach more efficient, safe, and insightful than existing industrial IoT techniques.

Performance measurements like latency, energy usage, and data throughput across deployment situations may help us examine this. Controlled trials and simulations in various industrial contexts may verify the model's efficiency and scalability. Using statistical techniques to compare the IoT-LPWM concept against centralized alternatives can quantify its energy savings and cybersecurity benefits. The model's generalizability across industrial applications is confirmed by this thorough research, which shows that the reported advantages are not context-specific.

4. Numerical results and discussion

With the advent of Industry 4.0, IIoT has become an essential technology. Wireless communication technology that is well-suited to industrial settings is LPWAN, which can cover a vast area with little power usage. This study has focused on learning to make the most of the IIoT network, which consists of network nodes and edge devices. An analysis of relevant studies revealed that edge devices and LPWAN had substantial computation and data communication capabilities. The suggested IOT-LPWM architecture takes advantage of all nodes in the network to provide useful data and insights. The interchange of information and knowledge map analysis between gateways and edge devices enhances the system's performance. Knowledge Map Analysis (KMA) gained from the suggested IoT-LPWM technique enhanced smart sensors and gateways, and data was gathered for various uses. Models were built in the cloud using machine learning and deep learning data analytics techniques. The findings aid pertinent experts in comprehending the system and predictively keeping the vital component.

The IoT-LPWM model's logical assessment criteria are based on measurements examining its industrial effectiveness. Real-time performance evaluation and data processing depend on latency. Integrating edge computing with LPWAN reduces energy usage, and Energy Consumption quantifies the model's efficiency. Data Throughput evaluates the network's data handling and transmission capability to ensure reliable operation. Finally, computing Complexity measures how successfully the model distributes processing workloads between edge devices and central servers, reducing central system computing load. These criteria objectively evaluate the model's performance, efficiency, and scalability.

Dataset Description:

Machine-learning intrusion detection systems may train their models in a centralized or federated manner. Edge-IIoTset, a realistic IoT and IIoT cyber security dataset, addresses this. To address IIoT and IoT demands, new technologies are suggested at each tier. After processing and analyzing the realistic cyber security dataset, do a major exploratory data analysis and evaluate ML approaches in federated and centralized learning.²⁵

i) Energy Management System (%)

Figure 5 deliberates the energy management ratio based on equation 1. In this era of Industry 4.0, the IoT is one of the essential technologies. The current crop of edge devices has powerful computational capabilities that can handle complex data sets. Because of its large coverage area and low power consumption, LPWAN is an ideal wireless communication technology for industrial settings. This article investigates the IIoT network's entire potential, comprising network and edge devices. A data and knowledge architecture was proposed after reviewing pertinent research demonstrating the significant processing and data-communication capabilities of LPWAN and edge devices. All nodes in the network would contribute to the development of useful data and insights under the proposed design. Interactions between gateways and edge devices allow data and knowledge exchange, improving the system's performance. Data is gathered for various reasons, and the information gained from the Internet of Things lightweight power wavelength (LPWM) strategy is used to enhance the functionality of smart sensors and gateways.

ii) Data Transmission Rate (%)

Figure 5.

Energy management ratio (%).

This condition monitoring and predictive maintenance component allows frequent time-based or event-based assessments of data obtained from assets’ sensors. When considering the Condition Monitoring & Predictive Maintenance component, it is necessary to analyze sensed data streams. One of the primary goals here is to determine the current status of technical equipment by constantly monitoring data streams from several sensors. Things like unusual actions taken by technological devices and the subsequent categorization of their current situation could fall under this category. The typical workflow aids with data collecting, pre-processing, AI technique setup, and post-processing.

Figure 6 explores the data transmission rate using equation 6. The user may insert nodes and connect them using edges. With smart devices, it is feasible to link the condition monitoring and predictive maintenance components to possibilities for heterogeneous data collecting. This way, a gadget may send out sensory data that has already been processed or is of a certain quality. Sensor data output is configuration-dependent and must be matched with an analysis flow. The edge device's data is in JSON format and it is possible to receive JSON configuration data, modify it, store it, and then send it back to the edge device. The edge device will reconfigure and transmit the updated sensory data using the specified parameters when pushed. While the administration view does let users conduct the JSON configuration, it does not understand or map its information. This method alleviates data integration conflicts, allowing diverse heterogeneous edge devices to be maintained.

iii) Performance Ratio (%)

Figure 6.

Data transmission rate (%).

Figure 7.

Performance ratio (%).

Figure 7 shows the performance ratio (%) through equation 5. An assessment of the current production status is a part of the planned Industrial Internet of Things system for predictive maintenance and condition monitoring. The cloud platform accepts data from various field devices via a high-performance gateway. Specifically intended to assist and allow predictive maintenance and ease the production status assessment, the core system board provides various instruments such as vibration, acceleration, and gyroscope information. To gain access to the intelligent data before it is circulated to the platform, the device performs a transformation in the frequency domain on the data that has been acquired. The MQTT protocol is used to communicate between the device and the platform.

iv) Attack Prevention Ratio (%)

The industrial knowledge map offers a graphical and sharable design for depicting manufacturing information. It benefits the integration and exchange of production information to enable intelligent applications. Collaborative construction of the knowledge graph is beneficial for gathering knowledge from various sources, including current data models, existing technical publications, websites, human experts, historical application records, etc.

Figure 8 expresses the Attack Prevention Ratio (%) based on Equation 2. This graph-based knowledge base lends a hand to the LPWAN's edge computing technologies, facilitating the distribution of heavy computing loads to the network's periphery and the semantic matching of industrial services. This allows for the integration of the computing resources of edge devices, which in turn reduces the computational complexity in the central. Not only does it lessen the energy used in processing and storing large amounts of data, it also lessens the likelihood of cyberattacks. In addition, the framework that has been developed allows for the discovery and generation of information and knowledge from many components of the system. These components include smart sensors, smart gateways, and the cloud. With this structure, it is possible to construct a ubiquitous knowledge network to enhance all the devices that are part of the system.

Figure 8.

Attack prevention ratio (%).

Experimental findings show that the suggested IoT-LPWM approach improves industrial monitoring and reduces energy usage. LPWAN and edge computing successfully optimized industrial wireless communication, reducing computational complexity. Clear and actionable insights from Visual Knowledge Map Analysis improved decision-making. Energy efficiency was achieved by shifting high-intensity processing to edge devices, reducing data processing and storage energy. Due to its decentralized nature, the IoT-LPWM paradigm lowered cyberattack risk and improved industrial communication network security. These findings demonstrate that the IoT-LPWM model may educate decision-makers and dramatically reduce industrial energy use. This validation underscores the model's pioneering position in industrial IoT, providing a holistic approach to energy-efficient and secure factory monitoring.

The IoT-LPWM model's flexible design allows experimental findings to be used in various industrial settings, independent of size or needs. LPWAN's broad coverage and dependable connectivity are vital in many production environments. The edge computing component may be tweaked to use the network's peripheral devices’ processing capacity to offload computationally heavy activities. Visual Knowledge Map Analysis's data-agnostic nature makes it ideal for evaluating many industrial data sources. The model's cybersecurity and energy-saving solutions may benefit any industry. The IoT-LPWM architecture's scalability and flexibility should allow other industrial settings to repeat the trials’ excellent outcomes, assuming the underlying infrastructure supports edge computing and LPWAN integration.

5. Complexity of the analysis

The integration of LPWAN and edge computing complicates the IoT-LPWM paradigm, requiring extensive network components and edge device synchronization. The paradigm moves high-intensity computing workloads to the edge, requiring durable edge devices with enough processing and storage. This dispersion decreases central server demand but complicates data flow management and real-time processing over a dispersed network. Using Visual Knowledge Map Analysis involves sophisticated data analytics and visualization technologies, significantly complicating the system. The approach is technically difficult yet efficient and secure due to the necessity for LPWAN and edge computing infrastructure interoperability and cybersecurity.

6. Conclusion

The Internet of Industrialized Things (IIoT) is becoming a game-changer thanks to the rise of Industry 4.0. The low-power wide-area network (LPWAN) is perfect for factories and other industrial settings since it can cover much ground with less power. Maximizing the IIoT network's potential has been the primary goal of this research. The network comprises both nodes and edge devices. Research on LPWAN and edge devices found they could communicate and compute data. The suggested IOT-LPWM architecture maximizes the network's performance by fully utilising each node. The overall system performance is improved when data is shared and analyzed via knowledge maps between gateways and edge devices. The proposed IoT-LPWM method improved smart sensors and gateways, and data is collected for several purposes using Knowledge Map Analysis (KMA). State-of-the-art data analytics techniques like deep learning and machine learning were used to construct working models in the cloud.

To sum up, this approach is data-driven, which means it has room for improvement to support a wider range of industrial operations than those covered here. The experimental results show that our proposed method IoT-LPWM achieves a high performance ratio of 97%, attack prevention ratio of 96.3%, energy management ratio of 93.8%, and data transmission ratio of 98.1% compared to other methods. The IoT-LPWM approach has issues but shows promise. A key drawback is the dependency on powerful edge devices, which may not be accessible in all industrial settings. Professional expertise and significant upfront costs may be needed to set up and integrate LPWAN with existing systems. The model's applicability and scalability to industrial settings must be improved in future studies.

Footnotes

Funding

This work was supported by (Research on Digital Transformation Construction Guide for Future Factory of Cigarette Production (ZJZY2023F004)

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

References

Flores-García

Jeong

Liu

, et al. Enabling industrial internet of things-based digital servitization in smart production logistics. International Journal of Production Research 2023; 61: 3884–3909.

Tyagi

Dananjayan

Agarwal

, et al. Blockchain—internet of things applications: opportunities and challenges for industry 4.0 and society 5.0. Sensors 2023; 23: 947.

Peter

Pradhan

Mbohwa

. Industrial internet of things (IIoT): opportunities, challenges, and requirements in manufacturing businesses in emerging economies. Procedia Computer Science 2023; 217: 856–865.

Sasikumar

Vairavasundaram

Kotecha

, et al. Blockchain-based trust mechanism for digital twin empowered Industrial Internet of Things. Future Generation Computer Systems 2023; 141: 16–27.

Yazdinejad

Zolfaghari

Dehghantanha

, et al. Accurate threat hunting in industrial internet of things edge devices. Digital Communications and Networks 2023; 9: 1123–1130.

Yazdinejad

Kazemi

Parizi

, et al. An ensemble deep learning model for cyber threat hunting in industrial internet of things. Digital Communications and Networks 2023; 9: 101–110.

Alanazi

Aljuhani

. Anomaly Detection for Industrial Internet of Things Cyberattacks. Computer Systems Science & Engineering 2023; 44.

Khan

Laghari

, et al. The collaborative role of blockchain, artificial intelligence, and industrial internet of things in digitalization of small and medium-size enterprises. Scientific Reports 2023; 13: 1656.

Wang

Lin

, et al. A distributed sensor system based on cloud-edge-End network for industrial internet of things. Future Internet 2023; 15: 171.

10.

Jahromi

Karimipour

Dehghantanha

. An ensemble deep federated learning cyber-threat hunting model for industrial internet of things. Computer Communications 2023; 198: 108–116.

11.

Awotunde

Folorunso

Imoize

, et al. An ensemble tree-based model for intrusion detection in industrial internet of things networks. Applied Sciences 2023; 13: 2479.

12.

Gunasekaran

Kumar

Kaladevi

, et al. Smart decision-making and communication strategy in industrial internet of things. IEEE Access 2023; 11: 28222–28235.

13.

Gopi

Sheeba

Anguraj

, et al. Intelligent Intrusion Detection System for Industrial Internet of Things Environment. Computer Systems Science & Engineering 2023; 44.

14.

Wang

Huang

Yin

, et al. Smart contract token-based privacy-preserving access control system for industrial internet of things. Digital Communications and Networks 2023; 9: 337–346.

15.

Agrawal

Nurmamatovna

Xudayarovich

, et al. Sustainable development with industry 4.0: a study with design, features and challenges. Journal of Integrated Science and Technology 2024; 12: 737–737.

16.

Rani

. Big data analytics challenges to implementing the intelligent Industrial Internet of Things (IIoT) systems in sustainable manufacturing operations. Technological Forecasting and Social Change 2023; 190: 122401.

17.

Lin

Zhou

Yang

, et al. Deep reinforcement learning-based task scheduling and resource allocation for NOMA-MEC in Industrial Internet of Things. Peer-to-Peer Networking and Applications 2023; 16: 170–188.

18.

Niu

Shao

, et al. Efficient heterogeneous signcryption scheme based on edge computing for industrial internet of things. Journal of Systems Architecture 2023; 136: 102836.

19.

Wang

Shang

Lei

, et al. Dual-attention assisted deep reinforcement learning algorithm for energy-efficient resource allocation in Industrial Internet of Things. Future Generation Computer Systems 2023; 142: 150–164.

20.

Niu

, et al. Attribute-based searchable encrypted scheme with edge computing for industrial internet of things. Journal of Systems Architecture 2023; 139: 102889.

21.

Khan

Alghamdi

. A neutrosophic WPM-based machine learning model for device trust in industrial internet of things. Journal of Ambient Intelligence and Humanized Computing 2023; 14: 3003–3017.

22.

Rashid

Khan

Eusufzai

, et al. A federated learning-based approach for improving intrusion detection in industrial internet of things networks. Network 2023; 3: 158–179.

23.

Singh

Gimekar

Venkatesan

. An efficient lightweight authentication scheme for human-centered industrial internet of things. International Journal of Communication Systems 2023; 36: e4189.

24.

Ali

Yousafzai

. Blockchain and federated learning-based intrusion detection approaches for edge-enabled industrial IoT networks: a survey. Ad Hoc Networks 2024; 152: 103320.

25.

Cheng

Sangaiah

, et al. Cognitive computing with a big data system in a secure internet of things. Applied Sciences 2023; 13: 7037.