Robert Playter: Boston Dynamics CEO on Humanoid and Legged Robotics | Lex Fridman Podcast #374

TL;DR

Boston Dynamics CEO, Robert Playter, discusses the development of humanoid robots, challenges in achieving natural movements, the importance of movement in robotics, and the future of robotics.

Transcript

• And so our goal was a natural-looking gait. It was surprisingly hard to get that to work. But we did build an early machine. We called it PETMAN prototype. It was the prototype before the PETMAN robot, and it had a really nice-looking gait where, you know, it would stick the leg out. It would do heel strike first before it rolled onto the toe, so... Read More

Key Insights

• 🤖 Achieving natural movements in humanoid robots is a complex task that requires years of research, simulation, and iterative development.
• ♻️ Underactuation and successfully managing complex environments pose significant challenges in legged locomotion.
• 😑 Robotics companies are exploring physical communication and expression to enhance human-robot interaction.

Questions & Answers

Q: What were the challenges in achieving natural-looking movements in humanoid robots?

One of the challenges was getting the robots to fully extend their legs and land on extended legs when walking. Underactuation, or the inability to push the robot in any direction, also posed challenges in achieving natural movements.

Q: How did Boston Dynamics simulate and model complex interactions and environments in their robotics research?

Boston Dynamics developed physics-based simulation tools to test and model interactions accurately, especially concerning foot-ground contact and manipulation of objects. The team focused on efficient and fast operation to simulate complex scenarios realistically.

Q: How did the company address communication and feedback between humans and robots?

Boston Dynamics acknowledges the tendency of people to anthropomorphize robots and embraces this by incorporating physical indications and expression in their machines. They are exploring ways to physically communicate the robot's intentions, emotions, or actions to improve human-robot interaction.

Q: What are the lessons learned from the DARPA Robotics Challenge?

The DARPA Robotics Challenge presented significant difficulties in accomplishing tasks such as driving vehicles, manipulating objects, and traversing complex terrain. The experience revealed the importance of building versatile, general-purpose robots that can adapt to uncertain situations and communicate their intentions and needs effectively.

Key Insights:

• Achieving natural movements in humanoid robots is a complex task that requires years of research, simulation, and iterative development.
• Underactuation and successfully managing complex environments pose significant challenges in legged locomotion.
• Robotics companies are exploring physical communication and expression to enhance human-robot interaction.
• General-purpose robots that can adapt to uncertain situations and possess physical indication capabilities will play an essential role in the future of robotics.

Summary

In this video, Robert Playter, CEO of Boston Dynamics, discusses the challenges and advancements in robotics, focusing on the company's humanoid robot, Atlas. He shares insights into the development of natural-looking gait and motion, the pursuit of elegance and beauty in movement, and the importance of simplifying complex robots. Playter also delves into the lessons he learned from working with Marc Raibert, the founder of Boston Dynamics, and the importance of pursuing one's interests and curiosity in robotics.

Questions & Answers

Q: When did Robert Playter fall in love with robotics?

Playter fell in love with robotics when he visited MIT and saw Marc Raibert's robots performing impressive movements. The connection between robotics and his interest in gymnastics sparked his fascination.

Q: What is the significance of elegance and beauty in movement?

In gymnastics, there is a concept of letting your body do what it naturally wants to do. Raibert aimed to build robots that could take advantage of this idea, allowing their physics to inherently guide their movements. The combination of natural physics and intentional control creates the potential for elegance and beauty in robotic movement.

Q: What were the distinctive lessons Playter learned while working at MIT's Leg Lab?

Playter learned the value of pursuing one's interests and curiosity, emphasizing the importance of doing something because you love it. This lesson has become a core value at Boston Dynamics.

Q: Why are roboticists often happy despite the challenges they face?

Roboticists are often happy because they have a strong connection to the physical world through their robots. The ability to see their creations moving in the world and working towards goals allows them to identify with their work, leading to a sense of fulfillment and happiness.

Q: What has Playter learned about life and robotics from Marc Raibert?

The most important lesson Playter learned from Raibert is to have the courage to pursue one's interests and curiosity, even if it may not seem financially rewarding. Raibert's fearless pursuit of challenging problems, such as legged locomotion, served as an inspiration.

Q: How tough is the effort of doing legged robotics, especially in the early days?

Legged robotics presented several challenges, including the need to balance and build a legged robot that could navigate effectively. In the early days, it required expertise in dynamics, feedback-control principles, and building real-time control algorithms that were able to handle the complexities of legged locomotion.

Q: Is dynamics more art or science in robotics?

Dynamics in robotics involves both science and art. Scientific principles are crucial in understanding and extrapolating legged locomotion, while the artistic side involves intuition and human-like movement. Humanoid robots provide a conduit for tapping into this intuition and expressing it in the machine.

Q: Can Playter share some memories from his time at MIT's Leg Lab?

Playter recalls the distinctive lesson he learned about pursuing interests and curiosity, and how it continues to be ingrained in Boston Dynamics' core values. The memories associated with the Leg Lab convey the importance of pursuing what you love and embracing your passions.

Q: Is it psychologically challenging to break robots during the development process?

Breaking robots is an essential part of the development process, as it allows researchers to understand weak points, make improvements, and iterate their designs. To overcome the fear of breaking robots, it helps to have the capability to repair and fix them, which boosts confidence and fosters a fearless attitude.

Q: Is there an art or science behind achieving elegant movement in robots?

Achieving elegant movement in robots involves both art and science. The understanding of physics, feedback control, and principles of motion serve as the scientific foundation. Intuition, observation, and the ability to perceive and capture lifelike movements contribute to the artistic aspect of achieving elegance in robot motion.

Q: How hard is it to get the humanoid robot Atlas to perform natural-looking walking and running?

Developing natural-looking walking and running for Atlas proved to be challenging over the years. Even with the PETMAN prototype, it was difficult to fully extend the leg and achieve a natural gait. However, the introduction of new techniques and advanced control algorithms enabled the team to achieve natural-looking motion in recent years.

Takeaways

Developing natural and elegant movement in robots is a challenging task that involves both art and science. It requires expertise in dynamics, feedback control, and simplifying complex systems. The pursuit of interests and curiosity, as well as the courage to tackle big problems, are crucial in the field of robotics. Breaking robots and learning from failures is an integral part of the process. General-purpose robots, like Atlas, can perform a wide range of tasks but face unique challenges due to their complexity. The development of robot behaviors has become faster with the advent of model-predictive control and optimization techniques.

Summary & Key Takeaways

• Boston Dynamics is a robotics company known for its creation of humanoid robots and physics-based simulation software.

• The development of natural-looking movements in humanoid robots took many years of research and trial and error, with the team achieving significant progress in recent years.

• Challenges included getting the robots to fully extend their legs, handle underactuation in movement, and manage complex environments and interactions.

• Playter highlights the importance of movement and elegance in robotics, as well as the integration of physics-based modeling and simulation in the development process.