Recently, Boston Dynamics posted a video of Atlas, Spot, and Handle dancing to Do You Love Me. To find out more about how this ties into the company’s approach to robotics for commercial applications, we caught up with Aaron Saunders, VP at Boston Dynamics, and asked him some questions. Here’s what he had to say:
Q: What inspired the idea of teaching robots to dance?
A: We were inspired by the idea of demonstrating robots in a fun way that could potentially open up new opportunities for robotic applications in our everyday lives.
Dancing offers an opportunity for us to see how robots can be programmed or taught complex movements that involve both agility and coordination. This is something that has never been done before and as such provides some unique insights into how we can use robots in commercial applications.
Q: How did you go about teaching Atlas, Spot, and Handle to dance?
A: First we used basic programming techniques like looping patterns of movements combined with trial and error until we got them right. Then we incorporated machine learning techniques using reinforcement learning algorithms so that the robots could self-learn certain movements over time based on rewards provided by our engineers.
This allowed us to get more complex movements quickly without needing too much input from our team. Finally, we built an AI-based motion control system that enabled us to create even more complex routines like those featured in the video.
Q: Can you tell us a bit about how this technology can help robotics development for commercial applications?
A: The technology behind our dancing robots has already been applied in several areas such as search and rescue operations where agility is key for navigating rough terrain entertainment where complex choreographies are needed manufacturing where precise repeatable motions are essential healthcare where remote manipulation is required logistics where autonomous navigation is necessary as well as many other fields.
By using computer vision technology combined with reinforcement learning algorithms and motion control systems these robots can be trained faster than ever before allowing companies to speed up the development process while improving accuracy and reliability at the same time.
Conclusion
Aaron Saunders has offered some valuable insights into the process of teaching robots to dance and how this technology can be applied in commercial applications. Through basic programming techniques, machine learning algorithms, and an AI-based motion control system, Boston Dynamics was able to create complex robotic routines with agility and coordination that have never been seen before.
This technology can help speed up robotics development while also improving accuracy and reliability across a variety of industries such as search and rescue operations, entertainment, manufacturing, healthcare logistics, and more. It is clear from Aaron’s comments that Boston Dynamics is paving the way for a new era of robotics innovation that will revolutionize the way we use robots in our daily lives.
Q&A with Aaron Saunders, VP at Boston Dynamics, on teaching robots to dance and how that informs the company’s approach to robotics for commercial applications (Evan Ackerman/IEEE Spectrum)
