Why Don’t We Have Better Robots Yet? | Ken Goldberg | TED

TL;DR

Creating robots for home use is challenging due to difficulties in hardware and software, including uncertainty in control, perception, and physics.

Transcript

I have a feeling most people in this room would like to have a robot at home. It'd be nice to be able to do the chores and take care of things. Where are these robots? What's taking so long? I mean, we have our tricorders, and we have satellites. We have laser beams. But where are the robots? (Laughter) I mean, OK, wait, we do have some robots in o... Read More

Key Insights

• 🤖 The reality of home robots is often far from the fantasy portrayed in media, with real robots often being clumsy and limited in functionality.
• 🤗 Hardware challenges, such as the need for reliable and simple hands, pose barriers to creating effective home robots.
• 🤖 Software challenges, including uncertainties in control, perception, and physics, make it difficult for robots to accurately and efficiently perform tasks.
• 🚚 E-commerce stands as a promising area for robots, where they can be utilized to sort and deliver packages efficiently, meeting the demand for timely delivery.
• 🤖 Ongoing research focuses on improving robot capabilities in manipulating deformable objects, folding laundry, and bagging items, among other tasks.
• 🤖 Despite the challenges, the speaker assures the audience that robots will eventually become more capable, but patience is needed in the development process.

Questions & Answers

Q: Why is it difficult to create home robots that can perform chores and tasks?

Home robots face challenges in both hardware and software, with the need for reliable and simple hands, as well as uncertainties in control, perception, and physics.

Q: Can you explain Moravec's paradox and its significance in robotics?

Moravec's paradox states that tasks that are easy for humans, like picking up objects and stacking blocks, are difficult for robots, while tasks that are easy for robots, like lifting heavy objects, are hard for humans. This paradox highlights the current limitations of robotics in mimicking human capabilities.

Q: How can AI and deep learning help robots improve their grasping abilities?

By training robots in simulation, using AI and deep learning algorithms, robots can learn how to grasp objects reliably without the need for explicit programming. This approach has shown success in picking up objects in e-commerce applications.

Q: What are some ongoing research areas to improve robot capabilities in homes?

Ongoing research aims to improve robot capabilities in manipulating deformable objects, such as strings and fabrics, as well as tasks like untangling knots, folding laundry, and bagging items.

Summary & Key Takeaways

• The gap between fiction and reality in home robots is explained, with real robots often being clumsy compared to the depictions in popular media.

• The hardware and software challenges of creating functional home robots are discussed, including the need for simple and reliable hands and uncertainty in control, perception, and physics.

• One area where robots have shown promise is in e-commerce, where they can be used to sort and deliver packages efficiently.