AI-Powered Robot Bicycle Mimics Batpod Stunts
Researchers develop an ultra-agile robot bike capable of wheelies, jumps, and front flips.
Engineers have created an artificial intelligence-powered robot bicycle, dubbed the Ultra Mobility Vehicle (UMV), that can perform dynamic stunts similar to those seen in Hollywood films.
The UMV, inspired by the agility of mountain bikers, balances on two wheels and can execute maneuvers such as wheelies, track stands, rear-wheel hops, and even front flips. Unlike conventional robots that rely on multiple wheels or legs for stability, this machine uses a powerful internal moving mass to shift its weight and control its momentum.
Developed by a team at the Robotics and AI Institute in Massachusetts, the 23.5kg UMV can travel at eight meters per second and jump over obstacles that are 130% of its height. Its design focuses on two key abilities: maximizing jump height and maintaining balance on a single wheel.
The researchers utilized computer simulations to test numerous designs before finalizing the UMV's bicycle-like structure. The robot stays upright not through external stabilizers, but by actively adjusting its weight distribution, with most of its mass concentrated in a movable 'head' unit above the frame. This configuration allows it to manipulate its center of gravity dynamically, much like a human rider performing tricks.
Training for the UMV involved reinforcement learning in a virtual environment, where the AI practiced thousands of simulated maneuvers. This trial-and-error process enabled the robot to successfully transfer its learned skills to the physical prototype. During testing, the UMV demonstrated a range of behaviors, including riding forward and backward, balancing stationary, hopping sideways on one wheel, and turning in place. A notable emergent behavior was a 'shimmy turn,' which developed organically from the AI's learning process rather than being explicitly programmed.
The UMV also autonomously navigated obstacles, successfully jumping onto a one-meter-high platform, traversing it, and dismounting. In another test, it completed a full front flip in under half a second before landing and continuing its ride.
Future applications for this technology could include urban delivery services, reconnaissance missions, or traversing challenging terrains where conventional vehicles would be impractical. The researchers noted that the UMV's combination of speed and compact size makes it suitable for such diverse roles.