2023-09-05
What will robots be like in the future?
Recently, Marc Raibert, founder of Boston Dynamics, revealed his vision for the future of robots: robots can do more than just dance
When Marc Raibert founded Boston Dynamics in 1992, he was not even sure if the company could become a robotics company. Now, Boston Dynamics has successfully launched Atlas bipedal robots and Spot quadruped robots, becoming the absolute authority of legged robots.
01 Shift from practical technology to strategic vision
Raibert, who comes from a technical background, has become more interested in pursuing his long-term vision of robotics technology as the company gradually expands in size and commercializes its territory.
To this end, Raibert established the Boston Institute of Power Artificial Intelligence in August 2022. The research institute is funded by Hyundai Automobile, and the first few projects focused on teaching robots to better understand the world around them, enabling them to function outside the laboratory.
Raibert hopes to teach robots to watch humans perform tasks, understand what they see, and then do it on their own, or know when they don't understand certain content, and how to ask questions to fill these gaps; Another goal of Raibert is to teach robots to inspect equipment to determine if it is working properly, and if not, to identify the problem and perform repairs.
These ideas represent Raibert's forward-looking vision for the future of robots. By endowing robots with the ability to "observe and learn", they can overcome the limitations of strict programming and achieve more autonomous and flexible operations. This requires the establishment of a powerful computer vision and knowledge representation system, which is an important step towards achieving universal artificial intelligence.
At the same time, if robots can autonomously detect and repair faults, it will significantly reduce maintenance costs and achieve higher levels of automation. However, this requires a deep understanding of the working principle of the equipment, and there is a large amount of abnormal data accumulation, which also poses challenges in terms of security.
In short, Raibert hopes to push robotics technology to a new stage and achieve universal applications.
02 Five Issues on the Next Development of Robot Technology
Question 1: Why do I share my vision for future robotics technology more with the outside world when I am in the research institute?
Leibert: At Boston Dynamics, I don't think we have discussed the vision. We just do the next thing, see how it goes, and then decide what to do next. I was taught that when you write a paper or give a speech, you should showcase the work you have completed, and what really matters is the data in your paper. Of course, you can talk about what you want to do, but people often talk about various things in this way... the future is so cheap and diverse.
This is different from showcasing what we have truly done at Boston Power. I am proud to showcase what we have done at Boston Dynamics. But if you want to establish a robot laboratory and you want to quickly complete it from scratch, you must outline a vision. So I'm starting to be more willing to do this, not to mention that we currently don't have any actual results to showcase.
Currently, robots must undergo repeated training to complete specific tasks. But Mark Leibert hopes that robots can watch humans perform tasks, understand what is happening, and then perform tasks themselves, whether in factories or at home.
Question 2: The research institute will invest a lot of energy in robots to better operate objects. What opportunities are there?
Rabert: I think people have been engaged in operational research for the past 50 years, but progress has not been enough.
I'm not criticizing anyone, but I think so much work has been done in path planning, but that's not the key. One idea is to shift from static robot operation to dynamic operation, which can advance this field just like shifting from static to dynamic foot robot landing technology.
Question 3: Do you worry about the public's negative perception of robots (especially those developed by oneself)?
Leibert: The media sometimes exaggerates stories about people's fear of robots.
I think overall, people do really like robots. Or at least, many people may like them, even if they are sometimes afraid of them, but I think people only need to know more about robots.
Question 4: What do you find interesting about dance robots?
Rabert: I think robots have many opportunities for emotional expression, but there is still a lot of unfinished work in this area. At present, creating these performances requires a lot of manual work, while robots do not perceive anything, they are only playing the behavior of our programming. But they should be listening to music, they should see who they are dancing with, and coordinate with it.
I must say that every time I think about this, I wonder if I have become weak, because robots possess emotions, both on the giving and receiving sides. But for some reason, it's fascinating.
Send your inquiry directly to us
