There are many systems at manufacturing sites that can only be operated by certain expert workers. Such a worker might rely on the subtle vibrations felt from the system to do extremely fine metalworking, or listen to slight changes in the sounds coming from a system to determine that it's beginning to fail. Manufacturing owes very much to the skills of these workers.
However, it is becoming increasingly difficult to continue manufacturing products with such a great reliance on their individual skills. In many countries and regions with aging populations, these experts are increasingly retiring from their careers—and younger workers seem disinclined to take their places.
Seeing how experts actually work to create systems for aging societies
In order to stay competitive, companies have begun using IoT systems at manufacturing and infrastructure inspection sites to search for ways to more efficiently transfer skills from experts to younger workers.
Image recognition technology can be used to detect and gather data from wearable sensors and videos of expert workers. Companies are using this data to see how experts work and have begun introducing systems that can reveal differences in how younger workers work. These systems are in fact already being used on some manufacturing lines. The way in which highly efficient experts work cannot be explained in words. However, these systems can be used to help train younger workers how to work at their level.
Some companies have taken things even further and have begun developing artificial intelligence (AI) on how experts work, in order to create machines able to function at the level of expert workers. AI that makes use of deep learning learns similarly to how an expert develops their insight and skills through work. The AI is trained repeatedly through an enormous amount of data, allowing it to develop skills based on tacit knowledge and sharpened insight that could not be manually programmed. These systems are now being used in actual factories and other sites.
There are many benefits to providing systems with functions that operate at the level of expert workers. Although a human learns only though experience, a system can share what it has learned with multiple machines, all of which will be capable of the same level of performance. This can be used to develop systems with advanced skills, which can then be easily deployed to factories all over the world. Unlike humans, these machines can work without rest as they continue to improve their skills, without any individual differences in performance. They therefore have the potential to improve both productivity and quality.
Detecting and seeing how experts work
Many companies are developing and implementing systems to see how their experts work, as a part of their digital transformation (DX) efforts. This section provides some examples.
Let's first consider an actual example of a system used to see how experts work and help train younger workers.
One company in the manufacturing industry is now using AI to analyze video of workers on a production line, in order to reduce working hours and prevent mistakes. Work procedures are often changed at this factory, as the company is involved in high-mix low-volume production. It is therefore crucial to do whatever it takes to prevent mistakes or drops in efficiency. The system used here analyzes video of workers to model how they work and move. It then compares this model with that of an expert worker to spot differences. This data is used to make daily improvements, allowing even new hires to quickly begin mimicking how experts move.
Metal mold fabrication requires especially fine machining, as workers must create precise metal molds using machines called machining centers. Some companies in this field have launched initiatives to analyze how these experts work. Normally, when detecting machining conditions based on data during the metalworking process, a current is passed through the metal (workpiece) and then the cutting resistance is measured based on the change in current while cutting. However, this method cannot be used to detect the subtle changes required to reproduce the skills of an expert worker. However, AI can now be used to analyze vibration data from the machine tool to detect machining conditions to a fine degree. Determining conditions required for high-quality metalworking in this way can help younger machinists to gain skills more quickly.
Automating high-level work that previously relied on experts
Our next example involves reproducing the skill of an expert in a system, as a solution to the shortage of workers in aging societies.
Some experts can determine what is wrong simply by sound when inspecting a failing machine that operates using a motor or engine or when checking for cracks in pipelines or leaks in walls. These workers use a special tool called an auscultation rod to hear sounds from specific points and are capable of astounding feats such as troubleshooting areas that cannot be seen based on how sound changes over several points. A system capable of this has now been developed. This system finds faults by using cloud-based AI to analyze data recorded through a microphone.
Similar technology is also being used in the tire industry. One tire manufacturer has developed a system capable of finely adjusting conditions during the production process—a task that previously had to be done by expert workers. Tires are made of rubber and become harder or softer with changes in air temperature. Production conditions must therefore be finely adjusted in order to reliably produce high-quality tires. Recent tires use much more complicated materials and structures, making them increasingly difficult to produce. The company had to rely on the insight and skills of expert workers with many years of experience to adjust production conditions. In order to make production easier, the company installed several hundreds of sensors on its production equipment to gather data on the state of rubber as it was being formed. They then used AI to determine the best method for adjusting conditions in real-time, and developed a system capable of adjusting production conditions, which they installed in their factory. They were able to increase productivity two times over, cut the number of workers down to one-third, and even improve the quality of their tires.
In order to develop systems with the skills of expert workers, a company must install sensors on the systems and equipment located in their factories or plants, in order to gather data and gain an accurate understanding of the situation. Having said that, gathering data from all these locations is not an easy task. There are often many limitations in place at work sites, such as where cables can be wired and where sensors can be attached. However, these issues are slowly being resolved. In December 2019, Japan introduced the ""local 5G"" system, in which companies other than communication providers can be licensed to install and operate 5G base stations in certain locations. This system can be used to build 5G networks in factories or plants, which will make it easier to gather a large amount of data at low latency from the many sensors installed on equipment and systems. Similar systems are being introduced in Germany and the rest of the world.
When an expert worker leaves, their skills will continue to be improved in the system
No matter how skilled an expert worker becomes, they will eventually leave the company one way or another—and that skill will be lost. However, the technologies introduced in this section can be used to develop systems that can continue to use these skills. What's more, such a system would continue to gain experience and learn even after the expert worker leaves, and its skills will continue to improve. If we can continue to inherit the technical skills of workers and the decision-making skills of managers, we might be able to solve the problem of passing skills on to younger workers as society continues to age.
You might be wondering what skills we should develop if these expert skills will increasingly be performed by systems. A machine still cannot come up with new skills on which it hasn't been trained. This means that the value of creative work will increase, regardless of the field.
- Continue reading:Expanding the scope of remote work using VR/AR and 5G technologies