Transforming industrial automation: voice recognition control via containerized PLC device

Abstract

The article discusses the impact of voice recognition and containerization technologies in the industrial sector, particularly on Programmable Logic Controller (PLC) devices. It highlights how voice assistants like Alexa, Siri, Cortana, and Google Assistant are pioneering and pushing the future of human–machine interfaces, with applications moving from smart homes to industrial automation. Containerization, illustrated by Docker, is transforming software deployment practices, offering benefits such as enhanced portability, modular architecture, and improved security when applied to industrial PLCs. The article introduces a novel approach to enhancing human–machine interfaces (HMIs) within industrial applications, leveraging voice recognition and containerization technologies on Programmable Logic Controllers (PLCs). Unlike traditional systems, this article integrates voice assistant with industrial PLCs through a containerized IoT architecture. This innovative framework enables efficient deployment on edge devices, supporting modular, portable, and secure operations aligned with Industry 4.0 and 5.0 paradigms. The study further includes a detailed implementation on microcontrollers and industrial PLCs, validating its application in a controlled laboratory environment and virtual model.

Introduction

Alexa¹, Siri², Cortana³, and Google Assistant⁴ have demonstrated what the future interface between humans and machines looks like. Voice commands have spread across smart devices, making life easier for millions of users by offering convenience. Along with industrial applications and smart home applications, voice assistants are also being added to driver assistance systems (according to Voicebot.ai, 73% of drivers will be using a voice assistant in their car by 2022⁵). Furthermore, voice assistants have found applications in healthcare and scientific laboratories. The market for voice assistants is continuously growing, and according to the research report by Market Research Future, "Voice Assistant Market - Information by Technology, Hardware, and Applications - Forecast to 2025", the market valuation stood at USD 1.68 billion in 2019. It is estimated to reach USD 7.30 billion by 2025⁶.

In recent years, containerization⁷ has emerged as a transformative technology, revolutionizing the way software applications are developed, deployed, and managed. Originally popularized by platforms like Docker, containerization provides a lightweight and portable solution for packaging, distributing, and running applications across various environments. While traditionally associated with web and cloud-based applications, the industrial sector is increasingly recognizing the benefits of extending containerization to programmable logic controller (PLC) devices.

PLCs play a critical role in industrial automation, controlling and monitoring machinery in manufacturing plants, power plants, and various other industrial settings. Traditionally, these devices operate on dedicated hardware and proprietary software stacks, making updates, maintenance, and scalability challenging. Containerization introduces a paradigm shift by encapsulating applications and their dependencies into self-contained units known as containers. This approach brings several advantages when applied to PLC devices in the industrial context.

One of the key benefits of containerization in the industrial sector is enhanced portability. Containers encapsulate not only the application but also its runtime, libraries, and dependencies (Fig. 1), ensuring consistent behavior across different environments. This portability facilitates seamless deployment of applications on various PLC devices, irrespective of the underlying hardware or operating system. As a result, industrial operations can achieve greater flexibility and agility in adapting to evolving technological landscapes.

Fig. 1.

Container encapsulates the application, runtime, libraries, and dependencies.

Furthermore, containerization promotes a modular and scalable architecture. By encapsulating individual functionalities or components within containers, industrial applications can be easily composed and decomposed, allowing for efficient scaling of resources. This modular approach facilitates the development of more agile and adaptable industrial systems, accommodating changes in production processes, equipment configurations, and business requirements.

Security is another critical aspect where containerization brings value to PLC devices. Containers provide a level of isolation between applications, reducing the risk of conflicts and vulnerabilities. Additionally, container orchestrators, such as Kubernetes⁸, offer robust security features, allowing industrial operators to manage and monitor applications effectively. This becomes increasingly important as industries strive to maintain the integrity and confidentiality of their automation systems.

For critical manufacturing systems that could endanger human lives, it’s essential that artificial intelligence (AI) does not directly control production processes. AI decisions can sometimes be incorrect or even hazardous, particularly if the system’s security is compromised by malware or hacking. Without robust security protocols, AI systems are vulnerable to exploitation, which could lead to severe operational risks^9,10. Therefore, AI should only be used for simulations and predictive analyses, with any suggested changes thoroughly vetted by humans before implementation to ensure safety and reliability. In our current system, AI does not influence the manufacturing process. For future research, we aim to expand the architecture to include AI-driven optimization and simulations of manufacturing processes.

Deploying applications to Programmable Logic Controllers (PLCs) involves configuring industrial-grade hardware to operate with real-time responsiveness, reliability, and security. Traditionally, PLCs run fixed programs, but containerization is revolutionizing this by enabling modular, portable, and easily updatable applications on edge devices. With containerization (like Docker), applications can be standardized, allowing easy transfer across environments, minimizing downtime, and enhancing scalability. This flexibility is especially valuable as industries move toward Industry 4.0 and 5.0, where adaptive, AI-integrated processes are required. In this way, containerization supports future-oriented manufacturing with better resource management and responsiveness¹⁰.

Cloud computing (Fig. 2) refers to centralized data processing in the cloud, which contrasts with Edge computing. Edge computing processes data close to its source, eliminating the need to send data to distant cloud systems, improving system performance and response times by reducing latency. The primary motivation for Edge computing is to bring AI, processing, and data analysis closer to sensors and actuators. Currently, architectures exist that use a combination of Cloud and Edge computing, creating a hybrid system that strategically distributes computational power across the architecture based on specific needs.

Fig. 2.

Cloud computing versus Edge Computing.

The concept of IoT architecture (Fig. 3) is straightforward: physical devices exchange data or control one another. Devices like smartwatches, light bulbs, refrigerators, cars, and buildings are part of this network. A key element in IoT is real-time data stream processing, where data is analyzed and used to make immediate decisions. Unlike traditional methods that store data before analyzing, IoT leverages data flows for instant insights. Advanced analytics identify patterns in real time, allowing connected systems to adjust settings or trigger alerts, integrating Cloud and Edge computing for optimized data handling (Fig. 3)¹¹.

Fig. 3.

IoT architecture.

In this evolving landscape, the implementation of containerization in PLC devices presents an opportunity for industrial enterprises to modernize their automation infrastructure, streamline maintenance processes, and improve overall operational efficiency. This journey towards containerized industrial automation represents a forward-looking approach that aligns with the broader trend of digital transformation in Industry 4.0¹².

The Industry 4.0 concept, globally accepted over the last decade, represents the digital transformation of industry through technology. Several countries have implemented their own strategies, such as Australia (Industry 4.0 Testlabs), France (Industrie de Futur), Japan (Society 5.0), and others. However, this technology-driven approach has not resolved social tensions or the climate crisis. In contrast, the Industry 5.0 concept emphasizes values and integrates the human factor into industrial transformation. It focuses on human comfort, health, and the working environment, striving to balance technological progress with human needs. Industry 5.0 is not a technological leap but an extension of Industry 4.0, emphasizing sustainability and regeneration. Climate change and biodiversity threats have revealed the inadequate resilience of globalization, as demonstrated during the COVID-19 pandemic. The world faces ecological challenges, making the transition to a circular economy essential. Industry 4.0, focused on economic growth, does not address these needs¹³:

Industry 5.0 encompasses three key aspects:

• Sustainability - Minimizing negative environmental impacts through the use of renewable resources, reducing emissions, and promoting recycling.

• Resilience - Creating flexible and adaptable production systems capable of facing crises and maintaining production continuity.

• Human-centricity - A synergistic working environment where people and intelligent machines collaborate, enhancing creativity and the safety of tasks.

Paradigms of Industry 5.0

• The European Commission defines Industry 5.0 as an evolution emphasizing long-term societal goals and the planet’s environmental limits. This concept includes:

• Social dimension - The psychological well-being of workers, their social inclusion, and technological solutions that complement human abilities.

• Environmental dimension - Supporting the elimination of fossil fuels, carbon capture, biodiversity restoration, and the circular economy.

Industry 5.0 represents a transformation that merges digital innovation with sustainability and resilience. By doing so, the industry prepares for future challenges and strives to achieve balanced and sustainable development. As we delve deeper into the integration of containerization with PLC devices, it becomes evident that the industrial sector stands to benefit significantly from the advantages offered by this innovative technology¹³.

Containerization technology and edge computing^14–16, when integrated with PLC devices, significantly enhance industrial automation systems. As already mentioned, containerization encapsulates applications, ensuring their seamless operation across varied environments. This is crucial for the distributed nature of edge computing, which processes data closer to its source, reducing latency and improving efficiency. Such a combination supports rapid deployment, scalable management of applications on PLC devices at the network’s edge, and fosters innovations across sectors by optimizing resource use and enhancing system reliability and security. As case study, we choose the integration of voice recognition directly on PLC devices through the containerized application, which is marked a pioneering advancement in industrial automation. Containerization empowers PLC devices by encapsulating voice recognition applications and their dependencies into portable and self-contained units. This approach allows for the seamless deployment of voice recognition functionalities on PLCs, transcending the constraints of traditional hardware and operating systems. By leveraging containers, the PLC devices become versatile platforms for voice-controlled automation, enhancing human–machine interaction in industrial settings. This innovation not only streamlines the integration of voice recognition capabilities into existing automation systems but also opens doors to more flexible and responsive industrial processes, aligning with the transformative possibilities of modern Industry 4.0 and 5.0^17,18 technologies.

Based on the mentioned reasons above, the article provides:

• Proposal of a new IoT architecture and method for controlling mechatronic devices using voice commands based on edge computing and containerization.

• Implementation of the proposed containerized IoT architecture on a microcomputer and industrial PLC device.

• Development of a software system for utilizing voice commands to control mechatronic devices in the IoT Edge environment.

• Verification of the proposed and implemented software system on a laboratory mechatronic device and virtual model.

System architecture

The uniqueness of this solution lies in the transformation of speech to text directly in the local network on IoT Edge devices, with the entire solution being containerized and easily distributable (Fig. 4). The chapter describes the use of Azure cloud¹⁹ and IoT Hub²⁰ for managing the chosen industrial PLC device, Revolution Pi²¹, which is an IoT Edge device. Three containers were created for voice control. The first, Voice Recognition Module, transforms speech to text. The remaining two convert text into specific commands and control connected devices. The Node-Red OPC UA module controls a designed water tank in Factory I/O²² software using the universal communication protocol OPC UA²³. The Node-Red OPC RevPi module is for controlling a real punching machine²⁴ from Fischertechnik. Both Node-Red modules offer control panels with available voice control commands and an overview of current events on the controlled devices via dashboard. The entire architecture is depicted in Figs. 4 and 5 and functions as follows:

• A.
  On the industrial PLC device RevPi Core 3²⁵, the initialization of the IoT Edge Runtime²⁶ occurs upon startup (Fig. 4 - point 1), which includes three basic modules that must run on the device for it to be considered an IoT Edge device:

  • IoT Edge Agent²⁷ represents a module responsible for deploying and managing other modules.
  • IoT Edge security daemon²⁸ ensures cryptographic operations.
  • IoT Edge Hub²⁹ serves as a local broker of messages between modules and the Azure cloud.
• B.
  After the successful initialization of the IoT Edge Runtime, the created modules are started (Fig. 4 - point 2):

  • Voice Recognition - this module is responsible for converting speech to text. For this conversion, the Azure Speech Recognition service³⁰ is used, which can run in the cloud or locally on a powerful IoT Edge device (Fig. 4 - point 3). The proposed architecture allows for both sound processing options, meaning both Edge and Cloud computing options. During initialization, the module attempts to connect to the Azure Speech Recognition service running locally (Fig. 4 - point 4). If this option is unsuccessful (due to service outage), the module automatically switches to connecting to the cloud (Fig. 4 - point 5), where the Azure Speech Recognition service also runs. The module also automatically recognizes the connected microphone and uses it as the sound source for voice commands. The module then sends the recognized speech text to the local message broker, IoT Edge Hub, which then forwards the messages to the other modules.
  • Node-Red RevPi - this module is used to control the physical punching machine with a conveyor belt (Fig. 6). Upon its startup, the initialization of individual input and output pins of the RevPi Core device occurs, which are necessary for controlling the punching machine (Fig. 4 - point 6). Successful initialization of the punching machine can be monitored through the lighting of white LED lights for motion sensors. After initialization, the module is ready to receive messages (speech converted to text) from the Voice Recognition module through the IoT Edge Hub module.

Fig. 4.

Architecture of the design of original voice control for the punching machine and water tank.

Fig. 5.

Flow diagram for the voice control of punching machine and water tank.

Fig. 6.

Punching machine with conveyor belt.

An important role in control is played by the connection between the Revolution Pi and the punching machine via the devices’ input and output peripherals, as shown in Table 1. In this study, Revolution Pi refers to the connection of RevPi Core and RevPi DIO. The values from various sensors on the punching machine, which are utilized within the proposed architecture, are detailed in Table 2.

Table 1.

Connection between the Revolution Pi and the punching machine via input and output device peripherals.

Function	Punching Machine	Revolution Pi
Power Supply (+) – Actuators	24V DC	+
Power Supply (+) – Sensors	24V DC	+
Power Supply (-)	0V	-
Power Supply (-)	0V	-
Phototransistor at the Start of Conveyor	I1	I3
Phototransistor Near Punch Tool	I2	I1
Upper Sensor – Punch Tool	I3	I2
Lower Sensor – Punch Tool	I4	I9
Conveyor Belt Forward	Q1	O3
Conveyor Belt Backward	Q2	O4
Punch Tool Up	Q3	O2
Punch Tool Down	Q4	O1

Table 2.

Connection between the Revolution Pi and the punching machine – sensor output values.

Sensor	Value	Meaning
Phototransistor at the Start of Conveyor	0	Object in field of view
Phototransistor at the Start of Conveyor	1	Object out of field of view
Phototransistor Near Punch Tool	0	Object in field of view
Phototransistor Near Punch Tool	1	Object out of field of view
Upper Sensor – Punch Tool	0	Punch tool is down
Upper Sensor – Punch Tool	1	Punch tool is up
Lower Sensor – Punch Tool	0	Punch tool is up
Lower Sensor – Punch Tool	1	Punch tool is down

The Node-Red RevPi module is tied to RevPi Core 3 as it utilizes its peripherals. The module is also responsible for converting text into specific control commands and offers a control panel accessible via a web browser. Created solution offers 6 control commands in English, and the entire punching process can be controlled through the control panel (Fig. 7). In Fig. 8, we can observe the Finite State Diagram for the Punching Machine. This diagram illustrates the different operational states of the machine.

Fig. 7.

Control panel for a punching machine with a conveyor belt.

Fig. 8.

State diagram for the Punching machine.

The control panel in Fig. 8 includes information such as:

• Recognized voice command or error message

• LED indicator signaling whether the punching process is ongoing or not. Red light indicates the ongoing punching process, and green light represents that the production line is ready for punching.

• Output values of sensors:

  • Object ready – signals whether the object is at the beginning of the belt and ready to start the process.
  • Object ready to punch - signals whether the object is in the punching position and ready.
  • Punching machine - signals whether the punching tool is up or down.
• The control panel also includes buttons for manual control of the process (Fig. 8), reflecting the following voice commands:

  • Hey Bennie, start punching process - the entire punching process is executed.
  • Hey Bennie, stop punching process - the punching process is immediately paused.
  • Hey Bennie, punching machine up - the punching tool moves upward.
  • Hey Bennie, punching machine down - the punching tool moves downward.
  • Hey Bennie, move belt forward - the conveyor belt moves forward.
  • Hey Bennie, move belt backward - the conveyor belt moves backward.

• Node-Red OPC UA - this module runs an OPC UA server, which is initialized upon module startup (Fig. 9). The OPC UA server then communicates with the OPC UA client on the side of Factory I/O - a virtual water tank (Fig. 4 - point 7). Once the module is successfully initialized, it can receive messages (speech converted to text) from the Voice Recognition module through the local message broker, IoT Edge Hub. The text is then converted by the module into specific commands for controlling the virtual water tank (Fig. 10). The module also offers a control panel accessible via a web browser. For the module to function correctly, it is necessary to properly configure the water tank in Factory I/O so that the client in Factory I/O can connect and read the correct variables from the OPC UA server running in the module (Fig. 11). The Node-Red OPC UA module is independent of the RevPi Core 3 device, as it does not utilize its peripherals. The control panel displays recognized voice commands and the tank fill level. The control panel also offers the option of manual tank control using buttons (Fig. 12). The water level in the tank is checked every second. The solution currently supports 6 voice commands in English:

  • Hey Bennie, open discharge valve - the discharge valve will be opened by 10% (predefined value).
  • Hey Bennie, open fill valve - the fill valve will be opened by 10% (predefined value).
  • Hey Bennie, close discharge valve - the discharge valve will be closed.
  • Hey Bennie, close fill valve - the fill valve will be closed.
  • Hey Bennie, open discharge valve by XX percent - the discharge valve will be opened by the value defined by us in percentage.
  • Hey Bennie, open fill valve by XX percent - the fill valve will be opened by the value defined by us in percentage.

Fig. 9.

Created Data Structure for the OPC UA Server in the Node-Red OPC UA Module.

Fig. 10.

Water tank designed in Factory I/O.

Fig. 11.

Configuration of Defined Variables for the Inputs/Outputs of the Water Tank via OPC UA.

Fig. 12.

Control panel for the water tank.

The Node-Red RevPi and Node-Red OPC UA modules run on our architecture on a common IoT Edge device for demonstration purposes of the capabilities of the RevPi Core 3. In real operation, the modules could run independently on two different IoT Edge devices.

In Fig. 13, we can observe the Finite State Diagram for the State Machine controlling the Water Tank. This diagram illustrates the various states of operation, such as Idle, Filling and Emptying.

• C.Successful initialization of the modules can be verified through the IoT Hub in the Azure cloud (Fig. 4 - point 8). If there is a problem with any of the modules, the error can be detected and investigated in the control center of our IoT Edge device. The IoT Hub in the Azure cloud can also be used for updating or deploying new modules. The created modules are stored in the Azure Container Registry running in the cloud, making it easy to distribute the modules to other devices or update already running modules.
• D.After the successful startup of all modules, the designed solution can be utilized. To activate voice control, it is always necessary to use the unique activation phrase "Hey Bennie," which activates voice control only when desired. If there was no activation phrase, voice control would be active continuously, causing constant computationally demanding transformation of speech to text and making it cumbersome to detect the start of commands, or a command detection error might occur. After a command is spoken, it travels through the microphone to the Voice Recognition module, where the transformation of speech to text occurs. The recognized text is then sent to the local message broker, which subsequently sends the text to Node-Red RevPi and Node-Red OPC UA, where it is the responsibility of the individual modules to process the text and convert it into specific commands.

Fig. 13.

State machine for the control of Water tank.

The main points of this architecture can be summarized as:

1. Speech to Text Transformation: The core innovation of this architecture is the transformation of speech to text directly within the local network on IoT Edge devices. This feature is pivotal for the voice command functionality.

2. Containerization and Distributability: The entire solution is containerized, making it easily distributable. This approach enhances the scalability and deployment efficiency of the system.

3. Azure Cloud and IoT Hub Utilization: The architecture leverages Azure cloud services and IoT Hub for managing the selected industrial PLC device, Revolution Pi, which is an IoT Edge device.

4. Container Modules for Voice Control:

5. The first container, the Voice Recognition Module, is dedicated to transforming speech into text.

6. The second container is the Node-Red OPC UA Module, designed for controlling a water tank in Factory I/O software using the OPC UA communication protocol.

7. The third container, the Node-Red OPC RevPi Module, controls a real punching machine from Fischertechnik.

8. Control Panels with Voice Commands: Both Node-Red modules (OPC UA and OPC RevPi) feature control panels that provide available voice control commands and offer insights into the current status of the controlled devices.

Evaluation and results

The popularity of voice control for IoT and IIoT is growing, and in the future, the majority of companies will have to face this challenge. Voice control can be beneficial in situations where physical contact with devices is difficult or even impossible. Voice control also brings a certain level of convenience for end-users. Voice control has already found its place in smartphones, where it is commonly used, and it is slowly gaining ground in the industrial sector. Based on the defined criteria for the industry usage, an innovative architecture has been created, bringing benefits such as:

1. Innovation - Innovative architecture introducing containerization and voice processing to industrial PLC devices.

2. Stability and low latency – As seen in Fig. 14, the response time is shorter when voice-to-text transformation is performed directly on the IoT Edge device rather than in the Azure cloud. The voice transformation by the IoT Edge device took around 2.97 ms, which means approximately a 58.37% better response time compared to Azure cloud response. Even in the case where the Azure Speech Recognition service ran in the Azure cloud, the response time for voice-to-text transformation was around 5.9 ms, using an internet connection with a speed of 15 Mbps. The experiment testing took place on a schedule between morning and evening at 24 °C (4 days, 7 h). In total, 56 attempts were made for speech-to-text transformation using both Edge and Cloud computing.

3. Data Protection – Since voice processing is performed locally (Edge computing), the data does not leave our local network and the risk of misuse is reduced. It also decreases the risk of data theft as they do not need to be transmitted over the network to a remote cloud storage. However, the solution also offers the option of data processing in the cloud (Cloud computing), where data protection is ensured directly by Microsoft, thanks to various certifications for working with sensitive data.

4. Portability – The modules were developed using containerization technology, enabling their use on other platforms that utilize containers, such as AWS IoT Greengrass²⁸ or Siemens MindSphere²⁹. However, a slight adaptation within the communication structures between the modules is necessary to make the developed modules compatible with the mentioned solutions.

5. Distribution and Maintenance – The solution can be easily distributed to many other IoT Edge devices in a short amount of time. Module updates are performed remotely through the Azure cloud. There is no need for physical access to the devices. Module deployment has always proceeded smoothly, taking approximately 1 min when using an internet connection with a speed of 15 Mbps.

6. Easy Usage – All created modules require no input from the end user. All initial configurations are performed automatically; for example, a connected headset is automatically recognized and utilized as a voice source.

7. OPC UA Protocol - By using the OPC UA protocol in developed Node-Red OPC UA module, we achieve interoperability, which is crucial in modern industrial systems.

8. Easy module customization thanks to Node-Red technology – The Node-Red OPC UA and Node-Red RevPi modules are created using flow-based programming of Node-Red, which is easily understandable, allowing for the easy expansion of the application.

9. Migration to ARM64 architecture – The created modules are not directly tied to the ARM32 architecture and can be easily moved to ARM64 by changing the base Docker image from ARM32 to ARM64.

Fig. 14.

Comparison of response times between speech recognition performed in the Azure cloud and on the Revolution Pi IoT Edge device.

The presented results are based on the design and validation of original procedures and methods for controlling mechatronic devices. These procedures and methods leverage IoT Edge, edge computing in conjunction with containerization, and Azure cloud, with a focus on industrial PLC devices. For these reasons, the article proposes a system that combines the advantages of voice control as a user interface and edge computing focused on personal data security, low latency, easy management, and application deployment through containerization.

A significant outcome of the work is the design and development of an application modular system with a modern HMI interface utilizing operator voice commands, complementing and enhancing conventional methods of controlling mechatronic devices. The new application modular system brings speed, simplicity in management, and deployment of developed applications. The described concept has been successfully tested and implemented. This comprehensive solution can be considered a scientific and practical contribution of the dissertation to the development of the cybernetics field.

The scientific and practical contributions represent the characteristic features of original procedures and solutions in the field of digital technologies applied in Industry 4.0 and Industry 5.0 concepts. These contributions are generalizable and can be further modified for additional research and technical practice. Here are the benefits of the proposed solutions as they relate to Industry 4.0 and Industry 5.0 concepts:

1. Digital Transformation: Industry 4.0 represents the digital transformation of industries through advanced technologies such as IoT, AI, and big data analytics​​.

2. Enhanced Automation: The integration of edge computing and containerization with PLC devices enhances industrial automation by ensuring seamless operation across varied environments, reducing latency, and improving efficiency​​.

3. Scalability and Modularity: Containerization promotes a modular and scalable architecture, allowing for the efficient scaling of resources and adaptation to changes in production processes and business requirements​​.

4. Improved Operational Efficiency: Modernizing automation infrastructure with containerized industrial systems streamlines maintenance processes and improves overall operational efficiency.

Benefits of Industry 5.0

1. Sustainability: Industry 5.0 emphasizes minimizing negative environmental impacts through the use of renewable resources, reducing emissions, and promoting recycling​​.

2. Resilience: This concept focuses on creating flexible and adaptable production systems capable of facing crises and maintaining production continuity​​.

3. Human-Centricity: Industry 5.0 aims to create a synergistic working environment where people and intelligent machines collaborate, enhancing creativity and the safety of tasks​​.

4. Social and Environmental Dimensions: The psychological well-being of workers, social inclusion, and technological solutions that complement human abilities are prioritized. Additionally, it supports the elimination of fossil fuels, carbon capture, biodiversity restoration, and the circular economy​​.

5. Enhanced Human–Machine Interaction: Innovations like integrating voice recognition on PLC devices through containerization enhance human–machine interaction, making industrial processes more flexible and responsive.

By combining these benefits, both Industry 4.0 and Industry 5.0 concepts aim to revolutionize industrial operations, balancing technological advancements with environmental sustainability and human well-being.

In Table 3, a comparative analysis of the proposed architecture with other available solutions highlights these limitations by evaluating each based on clearly defined parameters. This comparison illustrates where existing solutions fall short and establishes the necessity and unique advantages of the proposed architecture, thereby setting the foundation for this study’s contributions:

• Speech-to-text conversion performed using Edge computing technology

• Simple solution update and distribution across multiple devices

• Multilingual support

• Activation by phrase

• Utilization of containerization technology and deployment to industrial PLC

Table 3.

Comparison of the proposed architecture with other available solutions based on defined parameters.

Author	Speech-to-text conversion performed on IoT Edge device	Containerized solution Deployable to industrial PLC	Easy update and distribution of the solution to multiple devices	Multilingual support	Activation by phrase
Proposed and implemented architecture from article	Yes	Yes	Yes	Yes	Yes
Tran33	Yes	No	No	No	No
Li34	Yes	No	No	No	Yes
Tahseen Ali35	Yes	No	No	No	Not mentioned
Deuerlein36	No	No	No	No	Not mentioned
Aydogmus37	No	No	No	Yes	No
Nasef38	No	No	No	Yes	Not mentioned
Birch39	Yes	No	No	Not mentioned	Not mentioned

The contributions of the article are structured as follows:

1. Analysis of Containerization, Edge Computing, and Emerging Industrial Trends: This section explores how these technologies can be leveraged to design advanced methods for controlling mechatronic devices.

2. Design of a New IoT Architecture for Voice-Controlled Mechatronic Device Management: The proposed architecture utilizes voice commands, integrating edge computing and containerization to streamline control processes.

3. Implementation of the IoT Architecture on Microcontrollers and Industrial PLC Devices: This involves deploying the designed architecture on both microcontrollers and PLCs to ensure compatibility and functionality across different hardware.

4. Development of a Software System for Voice Command Control in IoT Edge Environments: A complete software solution is created to enable voice-controlled operation of mechatronic devices within the IoT Edge framework.

5. Verification of the Developed System on a Laboratory Mechatronic Device and Virtual Model: The proposed system is tested and validated in both lab settings and virtual simulations, ensuring effectiveness and reliability in real-world applications.

Author contributions

Conceptualization, L.B. and R.P.; methodology, L.B. and P.D.; software, L.B.; validation, L.B., P.D., E.K. and R.P.; resources, L.B., P.D., R.P. and E.K.; writing - original draft preparation, L.B.; writing - review and editing, L.B., R.P., P.D. and E.K.; supervision, E.K.; project administration, E.K.; funding acquisition, E.K. All authors have read and agreed to the published version of the manuscript.

Data availability

The datasets used and/or analyzed during the current study available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.