(Source: Lutsina Tatiana/Shutterstock.com)
When futurists of the past envisioned 2020, the vision likely included a plethora of smart devices that seemed at the time the exclusive domain of science fiction. They imagined such technology as domestic robots attending to household labor, auto-piloted cars and planes roaming the roads and skies, and immersive virtual reality as the next and perhaps final stage in audio-visual entertainment. Although this vision is not likely to manifest by 2020’s end because the world remains under the yoke of the COVID-19 epidemic, the Fourth Industrial Revolution (Industry 4.0) that embodies the vision remains underway.
As early as 2013, with the maturity of internet and computer technology and the gradual improvement of related infrastructure, Germany pioneered the concept of Industry 4.0—a new technological revolution that promised to leverage cyber-physical systems to improve people's lives in a multitude of different areas. This concept was later incorporated into the development plans of several additional countries, to create new areas of growth based on traditional industrial technologies and services by leveraging a combination of industrialization and informatization.
Today, the development and diffusion of technology related to Industry 4.0 is in full swing. On the software side, augmented reality technology can provide users with novel audio-visual experiences and has seen application in training programs for certain professions (such as police and doctors). Internet of Things (IoT) technology takes advantage of sensor clusters to achieve comprehensive monitoring of appliances. Also, advances in industrial cybersecurity technology allow timely monitoring of corporate networks to thwart hacker attacks. On the hardware side, 3D printing technology will enable users of any level to manufacture any design they can imagine quickly. The proliferation of industrial robots promises to standardize and streamline product manufacturing.
Data and machine-learning techniques that leverage data lie at the heart of this Fourth Industrial Revolution. Data is obtained from the sensors, transferred to a cloud server via the internet, and analyzed via machine-learning and artificial-intelligence algorithms. It is then returned to a service terminal or industrial robot to round out a complete workflow. Typical Industry 4.0 scenarios include building a better understanding of the user, product manufacturing, monitoring product quality, distribution logistics, and user feedback, each contingent upon the extensive involvement of data and machine learning.
A host of mobile phone and computer software programs already store and analyze user data. Some brick-and-mortar stores use radio frequency identification (RFID) chips to record user preferences and analyze user data via algorithms and other methods to recommend and update products and associated content. In a world run by Industry 4.0 user data—such as frequency of use, preferences, modes of use, and schedules—are recorded. This media runs the gamut from mobile apps to household appliances, office appliances, and medical devices. This data is analyzed via machine-learning algorithms to generate multi-dimensional classification labels. Each user is described by multiple labels, enabling an increasingly accurate portrait to be built for every user.
Comprehensive user profiles provide a very direct benefit in the form of increased personalization at the production level. Like how the content personalization is implemented for users who browse the internet today, in the age of the Fourth Industrial Revolution, highly refined user profiles will be directly applied to the product manufacturing process. This makes it easier for businesses to produce personalized products that meet users' needs. Personalized products can be tailor-made based on user data-driven predictions, invariably offering more possibilities to the user.
In addition to the potential impact on production decisions, control of the various steps in the manufacturing process will be fully automated with Industry 4.0 Internet of Everything (IoE) technology and industrial robotics. Each step in the production process will be fine-tuned in real-time based on an ongoing analysis of the outcomes of previous production steps and product requirements. In such smart factories, the controllability and robustness of production lines are improved, and worker involvement changes from repetitive work to the robotic agents’ supervision. Since its inception, Tesla has shown a commitment to building intelligent automotive factories, where not only production line assembly is performed by industrial robots, but warehousing, materials management, orders, and sales processes are all highly automated using AI, which has contributed to the company's industry-leading technology portfolio and sales.
In addition to the analysis and control of process-related data, a combination of machine-learning and machine-vision technology can automate large-scale, high-accuracy product inspections, which is particularly effective for identifying complex defects that cannot be easily ascertained by the human eye alone. Landing.AI, an artificial intelligence algorithm company led by renowned AI scientist Andrew Ng, recently launched an AI and machine-vision-based bubble detection device to detect gas leaks in devices. This machine-vision system allows the computer to capture small air bubbles with great precision and determine the gas leaks’ location. The error rate of the system's recognition routine is well below the 30 percent average error rate for inspections performed by workers finding bubbles by sight. When combined with data derived from the entire production process, this system rapidly locates problem areas and production lines and greatly reduces labor costs and identification errors.
At the end of the production process, provisions must also be made for the problem of logistics. Industrial robots can automatically package products and print specific QR code labels that include product information and mailing addresses on packages in preparation for their distribution. Autopilot systems are expected to play a major role in the distribution process. It is expected that in the next 10 to 15 years, automated driving technology based on computer vision, machine learning, and control technology will reach full commercialization, which will make delivery and logistics simpler and more efficient while significantly reducing human costs. The first batch of e-commerce giant Alibaba's smart robot warehouses was put into operation in 2017. Its subsidiary, Cainiao, has begun to implement technologies such as face recognition and drone-based dispatch. At the end of 2019, the valuation of Cainiao Logistics reached $28 billion (USD). In a future where the IoT and automated driving technology prevail, packages will seek out people rather than the other way around.
At the user end, data uploaded by a product's sensing system can be analyzed by machine-learning algorithms in the cloud to determine whether there are any anomalies in the data, enabling real-time monitoring of said product's performance. Furthermore, when a user encounters any issue, a trained AI system can efficiently handle tasks such as text chatting, answering phone calls, and video connectivity, allowing for rapid feedback and the timely resolution. The Bidirectional Encoder Representations from Transformers (BERT) model, published in 2018, has already surpassed humans in the field of chatbots, and related applications have taken on important roles in the product pipelines of internet giants (such as Microsoft Avatar, Ali Xiaomi, and IBM Watson) as well as emerging AI companies (such as Fourth Paradigm and C&T).
The distinguishing characteristics of Industry 4.0, which are demonstrated within the various applications mentioned above, primarily include the following:
Thanks to the Third Industrial Revolution (or Industry 3.0), people worldwide gained the ability to connect rapidly through the internet. However, in an Industry 4.0 world, sensors are integrated into each hardware piece to enable machine-to-machine communication. For example, in the printing and dyeing industry, a management system, acting as the center of the production system, coordinates raw dye allocation, dye positioning, automatic dispensing, automatic water provision, and proofing systems across the entire assembly line, achieving intelligent dyeing and greatly improving production efficiency and product stability. In addition, with machine-learning engines provided by information-physical systems and cloud computing, everything becomes truly interconnected—in other words, it becomes possible to seamlessly connect people to people, people to machines, machines to machines, and services to services. When connectivity becomes the norm, all factors in the process from production to service—including equipment, production lines, factories, and services—can be closely linked together.
Under an Industry 4.0 regime, the incorporation of information technology means that data necessarily becomes the lifeblood of industrial production. This data includes all aspects of production and associated services, including product data, equipment data, research and development data, supply chain data, operational data, and user data. On the one hand, data plays a decisive role in training and optimizing machine-learning algorithms. On the other hand, the deployment of machine-learning algorithms also requires constantly generated data so that the associated production processes can be properly controlled. This means that all aspects of life and production processes need to be digitized to the greatest extent possible. That is, everything must be quantified using reasonable metrics to embed automated systems effectively. This requires data scientists to design processes that take into account the current state of the data, to guide the system to collect the correct data consciously, and constantly design and optimize appropriate metrics.
As Industry 4.0 imposes relatively specific and detailed data flow requirements, various modules included in the production process will correspondingly become increasingly more refined. Each part of the production line becomes more and more modular and detailed, making personalized production possible while better reflecting and anticipating the users’ needs, creating a virtuous production-sales-feedback product cycle.
Industry 4.0 presents a plethora of opportunities. Although the entire production process can be integrated, the workload involved in data processes can be spread across multiple departments or even multiple companies. As a result, smaller companies’ single breakthroughs will become increasingly valuable as part of larger integrated processes. Similarly, intelligent devices can be divided using various classification, segmentation, and trend prediction models. They can also be separated into modules such as data transmission systems, data acquisition equipment, and data feedback systems, wherein each module can be embedded in other production processes. For example, data acquisition equipment can share production lines with other precision instrumentation production processes. Thus, various smaller enterprises can rely on Industry 4.0 and leverage their strengths to embed themselves in different market aspects.
The current status quo suggests that construction of infrastructure will become a key industry in the coming years. Both the friction between the U.S. and China that has occurred over 5G technology over the previous year and the fact that various internet companies have been building their cloud-computing platforms in recent years demonstrate the importance of infrastructure Industry 4.0 paradigm in terms of protecting corporate profits and national security. Also, data constitutes yet another form of infrastructure, and internet giants in possession of large amounts of data are most likely to enjoy the lion's share of opportunities. However, smaller companies will also have the opportunity to seek and identify segments of production and daily life that are not yet data-driven. Such opportunities will be particularly promising in industries where data is not yet available (such as the traditional heavy industry) and industries where data is not yet well used (such as healthcare).
The need for data and machine-learning technologies under Industry 4.0 is also a challenge faced by large mainstream corporations. Because large companies primarily rely on large-scale industrial production for their competitive advantage in our current development stage, the addition of sensing systems and IoT systems to their production lines will require a relatively large investment. Large companies also need to incorporate machine-learning technologies into production lines and product design, which requires the input of human talent and the innovation of management philosophy. The popularity of machine-learning technology in recent years has led companies to become all but obsessed with all things AI in their decision-making process, posing new challenges to decision-makers' ability to discern good investment from bad.
As Industry 4.0 gradually moves into the domain of people's everyday lives, many new applications that are impossible to predict today will be developed. When the IoT finally makes its way into millions more homes, and when automated driving is deployed on a large scale, humans will be freed from the large amounts of repetitive work they currently perform. Therefore, it is reasonable to ask: With the impact of Industry 4.0, will future careers be primarily concentrated in the computer industry or data analytics? Will people have more free time waiting to be filled? What will happen to interpersonal relationships as well as the relationship between man and machine? As we enter the third decade of the 21st century, these questions are still difficult for humanity to answer, but what is certain is that Industry 4.0 will constitute a major theme for the foreseeable future.
Wang Dongang is a PhD candidate in the University of Sydney. His research involves medical imaging, artificial intelligence, neuroscience and video analysis, and he is always devoting to implementing machine learning techniques into applications in daily life. He has published papers in top international conferences including CVPR and ECCV, and he serves as the reviewer for journals including IEEE Transactions on Circuits and Systems for Video Technology and IEEE Transactions on Multimedia and conferences including AAAI and ICML. He is experienced in developing algorithms in machine learning and computer vision. He has cooperated with companies and institutes in China, US and Australia in projects including multi-view action recognition, road management based on surveillance videos and auto-triage system for brain CT.
Privacy Centre |
Terms and Conditions
Copyright ©2022 Mouser Electronics, Inc.
Mouser® and Mouser Electronics® are trademarks of Mouser Electronics, Inc. in the U.S. and/or other countries.
All other trademarks are the property of their respective owners.
Corporate headquarters and logistics centre in Mansfield, Texas USA.