The Disney Research team unveiled a new child-size walking robot during their Wednesday presentation at the 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) in Detroit. The robot is related to the patent Disney just filed for a bipedal robot with different walking styles.
Evan Ackerman of IEEE Spectrum reported on the presentation, stating, “Disney Research has spent the last year developing a new system that leverages reinforcement learning to turn an animator’s vision into expressive motions that are robust enough to work almost anywhere, whether that’s a stage at IROS or a Disney theme park or a forest in Switzerland.”
The robot was developed by a team led by Mortiz Bächer at Disney Research in Zurich. The figure is mostly 3D printed, using modular hardware and actuators that made the creation process quick. The team was able to go from concept to this walking robot in less than a year.
Its head is able to look up, down, around, and tilt. Its legs have five degrees of movement, with hip joints that keep it balanced while walking.
Disney Research scientist Morgan Hope said, “Most roboticists are focused on getting their bipedal robots to reliably walk. At Disney, that might not be enough—our robots may have to strut, prance, sneak, trot, or meander to convey the emotion that we need them to.”
“What we try to bring to these kinds of robots is born from our history of character animation,” Michael Hopkins, a principal R&D engineer, said. “We have an animator embedded in our team, and together, we’re able to leverage their knowledge and our technical expertise to create the best performance we can.”
The researchers talked about the difficulties in bringing a potential animated character into the real world.
“In general, animation tools don’t have physics built into them,” said Bächer. “So that makes it hard for artists to design animations that will work in the real world.”
“It’s not just about walking,” said Pope. “Walking is one of the inputs to the reinforcement-learning system, but the other important input is how it walks.”
Ackerman writes, “Disney Research has developed a reinforcement learning-based pipeline that relies on simulation to combine and balance the vision of an animator with robust robotic motions.”
So an animator can create a highly expressive character, which the system then translates to real-life with the robot. Bächer states they have cut the time for Disney to develop new robots from years to months.
The system also ensures the robot will be able to handle anything it encounters in the real world, outside of test environments, while still emoting.
“This is a challenge for traditional techniques,” Ruben Grandia, an associate research scientist at Disney Research, said. “Normally, you have to hand-program this transition point. But if you put everything together in one simulation and perturb it while it tries to move and animate, it can determine that point for itself, which has resulted in recovery strategies that we see from this robot that we’d have no idea how to program.”
Disney has poured a lot of effort into developing robots and other systems that would allow a variety of characters to interact with guests at theme parks. At SXSW in March 2023, they demonstrated a small roller-skating robot resembling Judy Hopps. A true-to-life Hulk meet and greet was available for a limited time at Disney California Adventure last year, thanks to Project Exo. Disney has also tested life size Tinker Bell technology.
The key, Disney’s researchers explain, is to have the robot function as well as convey emotion.
“In situations where humans and robots are close to each other, conveying emotion and intent can be an important feature,” said Georg Wiedebach, senior R&D Imagineer. “So I think this can also be valuable in other applications where robots are working next to people.”
The team pointed out that while this specific robot is great, it’s really the development process that has been their big breakthrough.
“The idea is that this is a platform that’s hardware agnostic,” said Bächer. “So if we wanted to add more legs, or add arms, or make an entirely new character with a completely different morphology, we can rapidly teach it new behaviors. The off-the-shelf actuators, the 3D-printed components, our adaptable reinforcement-learning framework—these can all be applied to robots that are widely different in how they look and move. This robot is a promising first step on that journey.”
Disney scientist Espen Knoop said they will be pushing the limits of what’s possible with these robotic characters. “We want to see what happens when we get to those limits and learn what we can do at those limits.”