It is assessed by a humanoid robot with the help of an AI trained on recordings of human movements
Xuxin Cheng and Mazeyu Ji
An AI that helps humanoid robots mirror the movement of a person could allow robots to walk, dance and fight more convincingly.
The lightest and lightest robotic movements, such as from Boston Dynamics spectacular displays robot stunts are usually tight pre-programmed sequences. Teaching robots to perform a wider repertoire convincingly human movements are still difficult.
To overcome this obstacle, Xuanbin Peng At the University of California, San Diego, and his colleagues have developed an artificial intelligence system called ExBody2, which allows robots to copy different human movements and perform them more realistically.
Peng and his team first created a database of actions a humanoid robot might be able to perform, from simple movements such as standing or walking to more complex maneuvers such as difficult dance moves. The database included motion capture recordings of hundreds of human volunteers collected in previous research projects.
“Since humanoid robots share a similar physical structure with us, it makes sense to take advantage of the large amount of human movement data that is already available,” says Peng. “By learning to imitate these kinds of movements, the robot can quickly pick up a wide range of human-like behaviors. This means that whatever humans can do, the robot can potentially learn.”
To teach a simulated humanoid robot how to move, Peng and his team used reinforcement learning, in which an AI is given an example of what a successful movement is and then has to figure out how to do it through trial and error. First, they learned about full access to all the data of this ExBody2 virtual robot, such as the coordinates of each joint, in order to mimic human actions as much as possible. They then learned from these movements, but using only data they would have access to in the real world, such as measurements of inertia or velocity from sensors on a real robot’s body.
After training on the database, ExBody2 was put under the control of two commercial humanoid robots. She was able to smoothly connect simple movements, such as walking in a straight line and crouching, to more difficult movements, such as following a 40-second dance routine, throwing punches and waltzing with a human.
“Humanoid robots work best when they coordinate all their limbs and joints together,” says Peng. “Many tasks and movements require the arms, legs and trunk to work together, and full-body coordination greatly expands the robot’s range of capabilities.”
Topics:
