TOKYO -- In 2010, Rosen Diankov, a Bulgaria-born American, moved to Tokyo from the U.S. to pursue his lifelong dream of launching an industrial robot venture in one of the most robot-friendly countries in the world. He sees growing demand for his technology due to a shrinking workforce amid Japan's declining birthrate and the aging of its society.
With a Ph.D. in Robotics from Carnegie Mellon University under his belt, Diankov and his partner formed an industrial robot startup, called Mujin, in 2011. He has a big goal in mind: He aims to grow his robot business and help revitalize Japan through his innovative technology.
"If you ask me, the existing industrial robots have merely been manufacturing machines," Diankov said in fluent Japanese in Mujin's small office, about 105-sq. meters, near the University of Tokyo's main campus in central Tokyo. Most of the company's engineers are from various countries, and they talk to one other in English.
Diankov believes this because industrial robots would require dedicated operators to figure out ways to manually program each motion of robot joints, which is a time-consuming process. The existing robots simply repeat the programmed motions over and over again, and thus, they are unable to handle various tasks flexibly.
A new way
But Mujin's next-generation robot controller offers a solution and revolutionizes the way robots work. "Our Mujin controller is capable of detecting things in the surrounding area, and it creates real-time, autonomous motions accordingly," he said, adding that "95% of factory work needing automation can be controlled in real time."
For instance, the Mujin controller enables robotic arms to pick up target items from among a pile of parts in a box, and place them all neatly lined-up in a different box or fix them onto other parts. This seemingly easy task is difficult for conventional industrial robots to perform.
Mujin's proprietary 3-D vision system can identify a stack of parts, pick a target object and recognize its angle and height. Then, the Mujin controller will determine a collision-free path for the robot arm and send instructions to it.
Attaining such a high degree of flexibility is almost impossible with the conventional manual-labor method. "It would take at least one to two years to perform this picking task with the existing method. But our technology makes it possible to create such a program in one month, at the shortest," Diankov said.
This means Mujin technology can help manufacturers cut initial costs for introducing a new system. In some cases, he said, it can slash system development costs by half.
For years, these tasks have been handled manually by factory line workers. Yet, replacing them with industrial robots will greatly benefit manufacturers. Unlike humans, robots will never suffer fatigue from moving heavy objects, and they can operate 24 hours nonstop, a step toward achieving a "non-stop factory." Moreover, the use of robots can eliminate an inspection process aimed at preventing human error.
"It's got to be Japan"
At the Robotics Institute of Carnegie Mellon University, Diankov studied under Professor Takeo Kanade, who is an authority on robotics engineering. As such, Diankov had opportunities to meet a number of Japanese engineers who visited Kanade's laboratory. All these experiences helped lead him to starting a business in Japan, not in the U.S.
In addition, he sees potentially huge demand for his robot technology in Japan, because the country is facing a declining labor force and the hollowing-out of industry due to offshore production. "I've moved to Japan to commercialize my technology. It's got to be Japan," he said.
Unlike Japan, there is still strong resistance to greater use of industrial robots at U.S. manufacturers. American factory workers often voice opposition, saying that robots will take away their jobs. For Americans, improved efficiency means less work, which often results in job cuts. By contrast, many Japanese companies still maintain the lifetime employment system, which in turn ensures job security and employees' commitment to perpetual improvements, known as kaizen.
Diankov's interest in Japan traces all the way back to the days when he was studying as a freshman at the University of California, Berkeley. He was amazed by Sony PlayStation 2's highly sophisticated system configuration and parallel processing. That prompted his interest in learning the secrets of such high technological prowess in Japan. He then started reading Japanese-language books voraciously at the university library late into the night and mastered the language.
Only a week after he graduated from Carnegie Mellon University, he arrived in Japan in 2010, with a strong resolve to commercialize his technology. At first, he got a position as a postdoctoral researcher at the information system engineering laboratory at the University of Tokyo's Graduate School of Information Science and Technology, where he continued his research about how to market his technology.
Later, Diankov met his business partner Issei Takino, who now serves as Mujin's CEO, and succeeded in setting up Mujin in July 2011.
But the company got off to a rocky start. He received a number of harsh comments from Japanese manufacturers across the nation. They were skeptical about the Mujin controller's automation safety, and told him that they could not make a decision due to Mujin's lack of a proven track record. As Japanese manufacturers attach importance to technological safety and credibility, many of them were unwilling to try out Diankov's product. "Of course, safety is important. I have learned an important lesson," he said.
Yet, his perseverance finally paid off. He managed to launch a trial project with a customer who was open to the use of cutting-edge technology, and secured an order from his very first customer in 2012. Since then, Diankov has built achievements one by one.
Currently, Mujin has about 15 staff members. The company handles 10 to 20 projects a year at present, with plans to more than double its workforce within two years. To that end, it is moving into a bigger office by the end of August. "I will grow the company so that we can achieve sales of over 1 billion yen ($8.03 million) in four years," Diankov said.
"I hope to provide technology that will give a jolt to Japan's manufacturing industry. I will give a boost to our team as well as Japanese society as a whole," he added.