WATCH: Sneaker-wearing robot breaks speed record

Chinese robotics company, Robot Era, recently introduced Star1, the world’s fastest bipedal robot, capable of running at speeds up to 8 mph (12.98 km/h), with videos of it going viral on social media.

According to international media outlet, in November 2024 viral videos and photos of the robot, which sports sneakers, running across the Gobi Desert quickly gained popularity on Chinese social media, with many humorous reactions. However, few realised they were witnessing the fastest humanoid robot in action.

Star1, developed by Robot Era, is equipped with high-torque motors and AI-powered algorithms that allow it to navigate various terrains such as sand and grasslands. The robot’s efficient movement is aided by advanced motors, while its real-time processing abilities are supported by high-speed sensors and communication systems.

What sets Star1 apart from other bipedal robots in terms of speed is its use of human footwear. Standing 5.6 feet tall and weighing approximately 143 pounds, the viral Star1 can run at a speed of 3.6 meters per second, making it the fastest bipedal robot to date.

During testing, the robot, wearing sneakers, outpaced its barefoot counterpart, even though it started slower. The shoes helped the robot better navigate the challenging terrain of the Gobi Desert, ultimately allowing it to finish ahead.

Star1 currently holds the speed record, surpassing other advanced humanoid robots like Tesla’s Optimus, Boston Dynamics’ Atlas, and Unitree Robotics’ H1.

While these companies focused on advanced sensors and complex algorithms to improve robot motion, Robot Era achieved greater speed by optimising running mechanics and utilising existing technologies, like running shoes, to improve traction.

Back in November 2024, researchers from Johns Hopkins University unveils a successfully trained surgical robot using imitation learning, allowing it to perform surgical procedures as skilfully as experienced human surgeons.

This breakthrough eliminates the need to manually program robots for each specific movement during a medical procedure, moving the field of robotic surgery closer to true autonomy. In the future, robots could potentially perform complex surgeries independently.

The research team used imitation learning to teach the da Vinci Surgical System robot three fundamental tasks: handling a needle, lifting tissue, and stitching. By feeding the model hundreds of videos from surgeries performed by human doctors, the robot learned to replicate these tasks with remarkable precision.

Unlike traditional programming methods, which require hand-coding every step of a procedure, this approach enables the robot to learn from demonstration videos, significantly speeding up the process of training.

The key innovation is combining imitation learning with machine learning architecture similar to that of ChatGPT, but instead of processing text, the model works with kinematic data that translates the robot’s movements into mathematical calculations.

This allows the robot to predict the necessary movements based on camera input without needing explicit instructions for every action.

This breakthrough could drastically reduce the time required to train surgical robots, enabling them to learn a wide range of procedures in just days.