Massachusetts Researchers Developed Stunt-Performing Robot

The Ultra Mobility Vehicle uses AI to balance on two wheels and perform high-flying acrobatic jumps.

Updated on Oct. 14, 2026 in Robotics

Massachusetts Researchers Developed Stunt-Performing Robot

Live Poll

Would you trust an autonomous robot to deliver goods to your home?

Researchers in Massachusetts have developed an AI-powered robot bicycle capable of executing complex stunts, including jumps and flips. The Ultra Mobility Vehicle, inspired by trail riders, balances on two wheels while navigating obstacles with precision.

Why it matters

The development represents a significant advancement in autonomous mobility and dynamic control systems. By mimicking human mountain bike techniques, the robot demonstrates how AI can achieve high-performance agility in challenging physical environments.

The Ultra Mobility Vehicle weighs 23.5kg and reaches speeds of 8 metres per second. It can execute a front flip in less than half a second and clear obstacles 130 per cent of its own height.

The players

Ultra Mobility Vehicle

This is an AI-powered robot bicycle designed to perform stunts and navigate complex terrain.

Science Robotics

This is a peer-reviewed academic journal that publishes significant developments in the field of robotics research.

The details

Engineers trained the machine in virtual environments using reinforcement learning after testing dozens of designs in simulations. The robot achieves its acrobatic feats by utilizing a movable head to shift weight, allowing it to maintain balance on two wheels and jump onto platforms reaching one metre in height.

The design was published in Science Robotics on October 14, 2026.

The Tech Race

This development mirrors the rapid evolution of robotic agility seen in platforms like the Boston Dynamics Spot robot, which fundamentally shifted expectations for mobile hardware. The Ultra Mobility Vehicle moves the field beyond industrial traversal toward high-speed, acrobatic performance in urban environments.

Future iterations of this technology could eventually streamline urban delivery systems by navigating obstacles that current wheeled bots cannot traverse. Users may see these robots deployed for reconnaissance missions or automated logistics tasks as control algorithms improve.

The takeaway

The success of this robot highlights how virtual reinforcement learning can be used to translate complex human athletic skills into machine motion. Implementing similar balance-shifting technology could improve the stability and terrain-handling capabilities of consumer-grade autonomous devices.