Disney’s new robots are here! Without further ado, let’s first take a look at what it looks like – big twinkling eyes, a shaking head, flexibly swinging antennae, and a pair of short legs.



The walking is bumpy, but the key is still very stable:


Although it is a machine, I have to say that this expression is very "spiritual":


At the just-concluded international robotics conference IROS, Disney Research Institute research scientist Moritz Bächer led a team to demonstrate this new robot that wowed everyone.


Although this little guy doesn't have an official name yet, his performance is impressive enough.

The biggest difference from common bipedal robots is that it can not only maintain dynamic balance while walking, but also express some "emotions" through movements.

According to the researchers, this is mainly due to the use of the training process that the team spent a year developing last year.reinforcement learning system, and also hired professional animators to design movements, and applied the way animated characters express emotions through movements to the robot.

Based on this, the little guy is no longer a cold machine. It is conceivable that if a GPT brain is installed on it, the dog at home will be in danger~~~


Risibai, I am a college student, give it to me.

Many netizens thought of the characters in Star Wars after seeing it:


Another wave of netizens saw the little guy being so "naive" and thought he was the real-life version of Wall-E.(WALL-E, character in "Wall-E"):


Some netizens expressed that they felt like they were “standing on the cusp of the robot revolution”:


Why does this little guy walk so smoothly? How exactly do developers give it "emotion"? Let's look down.

"Walking" in Swiss forests

The secret to moving smoothly while shaking your head lies in the design of the head and legs.

Among them, the little guy’s head is afour degrees of freedomThe design enables tilted observation in four directions: up, down, left and right:


And its legs have increased tofive degrees of freedom,havehip joint, so it can maintain dynamic balance while walking.

Push it sneakily, or drag the carpet under its feet deliberately, and it can handle it with ease:



Challenge more complex terrains, including walking in the wild forests of Switzerland:


Steady yet charming, you can even dance to a line while walking:


According to researchers, it actually took less than a year to achieve this effect from concept to final polishing.

This is also due to3D printingtechnology, useModular hardware and actuators, greatly improving the speed of design and iteration.

Let’s talk about giving it “emotion”.

In fact, making robots move in an emotional way is Disney's specialty. As early as the 1960s and 1970s, Disney developed mechanical dolls.

But as robots become more advanced and mobile, designing emotional behaviors that are compatible with them becomes challenging.

To this end, Disney pulled out its trump card and hired a character who is good at making characters convey emotions through movement.animatorto design actions for the robot.

But when it comes to animators designing actions, they have to face a problem——

Animation tools often do not have built-in physics, and animators' visions are sometimes difficult to achieve in reality. This requires the team to spend a lot of time experimenting.

To bridge this gap, the team used a reinforcement learning system.

Behind it is the blessing of reinforcement learning

The research team developed a reinforcement learning-based process tosimulationCombine the animator's visual vision with the robot's movements.


This process essentially provides the animator withphysical constraints, ensuring that the highly expressive movements designed by the animators can be performed in reality, or as close as possible to what the robot can do.

Morgan Pope, a research scientist at Disney Research Institute, said:

This is not just walking. Walking is only one of the inputs to the reinforcement learning system. Another important input is how to walk.

Fortunately, in this case, one of the advantages of reinforcement learning is that the generated motion is very robust.

And according to the researchers, this process can train a robot to do new actions in just a few hours on a PC, greatly shortening the time required to develop new robot characters.

At the same time, during repeated training, subtle adjustments can be made to the motor performance, mass distribution, and friction between the robot and the ground.

Team leader Bächer believes that "this has shortened the time required for Disney to develop new robot characters from years to months."

And the researchers also emphasized that the important thing is not the robot, but the process:

So if we want to add more legs, arms, or make a whole new character, we can quickly teach it new moves. Ready-made actuators, 3D printed components, and adaptable reinforcement learning frameworks can all be applied to robots with widely different appearances and movements.

And this robot is just one step in that journey.

Reference links:

[1]https://twitter.com/linusekenstam/status/1710277661069574171

[2]https://spectrum.ieee.org/disney-robot