KAIST presents a robot that runs and dances the moonwalk: video.

According to Korrespondent.net: Researchers from the Korea Advanced Institute of Science and Technology (KAIST) have introduced a new humanoid robot that not only runs quickly and overcomes obstacles but also performs Michael Jackson's famous 'moonwalk' move, as reported by Interesting Engineering.

Outstanding capabilities of the new robot

At the start of the demonstration, the robot moves backward, performing smooth gliding steps on a conveyor platform. It then showcases various walking styles: from a calm pace of 0.98 m/s to running at a speed of 3.3 m/s (about 12 km/h). It was also tested for stability: even after jolts and impacts, it remained balanced and continued to move.

One of the key features is its ability for 'blind walking': the robot navigated obstacles, stairs, and debris relying solely on internal sensors, without the use of cameras or visual sensors. It also demonstrated 'duck walking' and jumping on straight legs, indicating a high level of accuracy in control algorithms.

Technical specifications of the robot

This humanoid stands 165 cm tall and weighs 75 kg, equipped with motors and reducers developed at KAIST. Its control system is based on artificial intelligence using reinforcement learning conducted in a simulator, which has reduced the gap between models and reality.

The team will present their research findings at the CoRL 2025 conference on September 29 and IEEE Humanoids 2025 on October 1. The scientists plan to teach the robot to perform more complex tasks such as moving with carts, climbing ladders, and working in industrial conditions.

Interestingly, recently Xpeng also introduced its humanoid robot that is already operational at its production facility.

The first football match between robots took place in Beijing

News from Korrespondent.net on Telegram and WhatsApp. Subscribe to our channels.

The development of such robots opens up new possibilities for their applications in various fields. From autonomous systems in manufacturing to interactive assistants in everyday life — they can significantly ease many processes. Further research in this area promises even more innovations that could change our perception of future technologies.