The Robotics and AI Institute, which develops advanced robots and intelligent machines based in Massachusetts, has raised the bar for cycling robots this year with freestyle tricks learned by AI. Moreover, it does so on one of the most exclusive bikes that a child can learn to ride.
The robot that does front flips… on a Specialized Hotwalk Carbon
Advertising
The Ultra Mobile Vehicle (UMV), the cycling robot developed by the RAI Institute, has once again shaken the internet. Just a few months ago, it amazed with impossible balances, jumps, and backward driving; now it has taken a step further and is capable of chaining complete front flips, a maneuver more typical of any MTB rider than a laboratory robot. The scene would already be striking in itself, but it is even more so when we see that the UMV continues to use a Specialized Hotwalk Carbon, a balance bike for children that shares some of the technology that Specialized mounts on its World Cup-winning XCO models.
The Hotwalk Carbon is a rarity within the children's segment.It cost around €600, is currently sold out and discontinued, and Specialized designed it as if it were a scaled-down version of its competition bikes, with a frame, fork, handlebars, and wheels made entirely of carbon, an ultralight FACT 9r frame, wheels with Rhythm Lite tires, a low-friction Body Geometry saddle with a handle for transport, and a proportional geometry that seeks to replicate the sensations of a high-performance bike. Overall, it is a surprisingly advanced platform for a robot… and for any child.
Beyond the spectacle, what makes the UMV truly unique is its way of learning. It is based on Reinforcement Learning, a system where the AI tests, fails, corrects, and tries again until it masters each gesture. It is the digital equivalent of a rider repeating a trick until it comes out clean. This method explains how it has gone from maintaining balance to executing a complete front flip with unexpected fluidity in such a short time.
Advertising
On its official page, the CEO of RAI Institute, Marc Raibert, summarizes the mission: “We need to make robots smarter, more agile, and skillful, and easier to use — more like people… Robots will increase productivity, free people from dangerous jobs, care for dependent individuals, and, in general, help us live better.”
An ambitious goal that contrasts with the current images of the UMV: a small robot smashing the boundary between balance and freestyle.
Training
Bike Equipment
Bike Equipment
Cyclocross
Cyclocross
Gravel