Skateboarding Robots: Teaching Machines to Master Physical Contact [English]

Author: Gadget Review | Source: Gadget Review | Read the full article in English

Researchers from the University of Michigan and Southern University of Science and Technology have made a significant breakthrough in robotics by developing a system that enables four-legged robots to skateboard. This new framework, known as DHAL, allows robots to perform complex movements that involve physical contact with objects, such as skateboards. Unlike previous methods that struggled with sudden changes in movement, DHAL helps robots transition smoothly between different actions, making it easier for them to learn and adapt.

The challenge with traditional robots is that they can walk and run well but often fail at tasks that require precise interactions with their surroundings. The researchers recognized this limitation and created a system that can handle these intricate movements without needing extensive programming. By using a method that autonomously identifies when a robot changes its movement phase, DHAL simplifies the learning process and enhances the robot's ability to perform tasks that mimic human actions.

While the idea of skateboarding robots may seem like a fun novelty, the implications of this technology are far-reaching. The researchers believe that robots equipped with this framework could be used for practical applications, such as delivering packages in urban areas or navigating office spaces more efficiently. As they continue to refine this technology, they aim to apply it to other complex tasks, potentially revolutionizing how robots interact with the world around them.

[Read More (translated)]