Wang Xingxing (Unitree) describes a bottleneck: robots are still less efficient than humans and take time to learn. Going from 80% to 99.9% success is the real test.
On the sidelines of the World Robot Conference in Beijing, industry leaders delivered a sober diagnosis: the main obstacle for humanoid robots remains to make the technology work properly. Robots are not yet as efficient as humans and take time to acquire new skills, which creates a bottleneck for the entire industry, explained Wang Xingxing, founder of Unitree — a day after an IPO whose stock jumped 460% on the first day, before falling 18.7% the following Thursday.
The same is true for deployment. Last month, the U.S. Federal Communications Commission added foreign-made advanced robotic devices, including humanoids, to a list restricting imports, without explicitly targeting a country. Jeff Burnstein, president of the Association for Advancing Automation, puts the immediate scope into perspective: few humanoids are actually used in the United States, and the effect would be felt more on autonomous mobile robots, which are much more present in factories. What his customers are asking for, he adds, are solutions to a problem — humanoid or not, robot or not.
Wan Bin, chief operating officer of the start-up Keenon, which combines humanoids and simpler delivery robots for hotel laundry services, puts the real difficulty at ease: completing 50 percent of a task correctly is relatively easy, 80 percent as well; it's 99.9 percent that tests engineering and training. The company, which counts Hilton and Buffalo Wild Wings among its partners, says it has delivered more than 100,000 robots and is aiming for more than 150,000 units by the end of next year.
Read the full article on the original source
Go to the source →