US seized robot leadership from Japan during Fukushima disaster

The United States took the lead in the robotics industry after Japan's robots failed to overcome the radioactive environment during the Fukushima nuclear disaster. KAIST Professor Gong Gyeong-cheol noted that while Japan specialized in rigid position control, the shift toward "Physical AI" requires flexible force control.

The axis of the robot industry is shifting. The paradigm is transitioning from the era of industrial robots that repeat sophisticated movements to the era of "Physical AI," where artificial intelligence understands physical environments and interacts smoothly. In this process, an analysis has emerged regarding why Japan, once considered a powerhouse in robotics, has seen its position shrink and where the new technological leadership is headed.

An interview with KAIST Professor Gong Gyeong-cheol of the Department of Mechanical Engineering, reported by T Times TV, contains the historical flow of robot technology, cultural backgrounds, and the technical capabilities required in the rapidly changing AI era. Professor Gong explained that the development of robot technology is closely related to the cultural icons that engineers encountered during their childhood.

Technology created by cultural romance, and the crisis of Japanese robots

Professor Gong analyzed that the tendencies of engineers are determined by the cultural works that influenced them. While a generation raised on The Terminator might focus on the threat of robots and ask, "Are you going to destroy the world?", a generation that watched Astro Boy perceives robots as friends. The Mazinger Z generation views robots as saviors to be deployed to disaster sites, while the generation fascinated by Baymax or soft robotics since the 2000s perceives robots as "soft and warm" beings, leading to the development of care robots or patient transfer robots.

This cultural background led to actual directions in technology development. Professor Gong specifically pointed to the Fukushima nuclear disaster as the decisive factor in the collapse of Japan's robotics stronghold. While the world's eyes were on robots expected to be deployed to disaster sites, Japan's robots failed to overcome the radioactive environment, allowing the United States to seize the leadership. Through the DARPA Robotics Challenge, the United States presented missions similar to the Fukushima nuclear disaster and preempted disaster recovery solutions, resulting in the leadership of robot brands shifting from Japan to the United States.

The fundamental reason Japan lost its technical leadership lies in the difference in control methods. Japan's actuators and control technologies were specialized in repeating specific movements very precisely and rigidly. However, in the AI era, the paradigm of robots has changed from rigid position control to "Force Control," which adjusts force smoothly to suit the environment. Professor Gong noted, "Japan's existing technology lacks the flexibility suited for the AI era."

Localization of materials and components and the platform strategy of Physical AI

Regarding whether South Korea has escaped its dependence on Japanese materials and components, Professor Gong asserted, "It has not." While it was once natural to think of Japan when mentioning robots or materials and components, the 2019 export restrictions imposed by Japan are evaluated as having been a "heaven-sent opportunity" for the South Korean robot industry. At that time, Korean companies began to internalize core components, including actuators, and have now reached a level where major components, such as winding the coils inside actuators, can be procured domestically.

The actuators and control architecture, which are core technologies of Angel Robotics, are expected to play an important role in the future "Physical AI" ecosystem. Professor Gong explained that actuators will go beyond simply acting as robot joints and will serve as basic infrastructure that helps AI resolve noise and nonlinearity in physical environments to obtain refined data. This is in a similar context to the role of NVIDIA, which provides GPUs and the CUDA engine to run AI models.

He predicted that the future robot industry would follow in the footsteps of the smartphone industry. Just as smartphones expanded their ecosystems through various apps, the Physical AI era must see the emergence of complex and diverse types of robots. For example, technology should permeate daily life to the point of being invisible, such as a microphone that moves on its own to listen to sound. For this, Professor Gong's analysis is that the role of "platforms," such as actuators, control algorithms, and data centers, is as important as the performance of individual robots.

In the AI era, talent with "distinctive expertise" will survive

What capabilities should future talent possess in a rapidly changing technological environment? Professor Gong emphasized specific expertise, such as "regional characteristics," rather than universal knowledge. This means that instead of general-purpose talent covering all fields, the ability to handle data and environments specialized for specific industries is essential.

For instance, expertise is needed to study Physical AI specialized for manufacturing environments or to obtain data from the agricultural field and utilize it for control. Professor Gong suggested, "Rather than unconditional universality, distinctive expertise that can obtain data from a specific field and connect it to physical control will be the core competitiveness of the future."

Source: original video (YouTube)