Real-life Batpod! Incredible robot bicycle performs wheelies, jumps and front flips like a Hollywood stunt bike

If Batman needed a new Batpod, engineers may have just built it.Researchers have created an AI-powered robot bicycle that can pull wheelies, leap over obstacles and even perform front flips, mimicking the gravity-defying stunts seen in the Dark Knight films.The machine, known as the Ultra Mobility Vehicle (UMV), was inspired by the tricks performed by mountain bikers and trail riders.Unlike conventional robots that rely on multiple wheels or legs for stability, the UMV balances on just two wheels.At the heart of the design is a powerful moving mass that allows the robot to shift its weight and control its momentum mid-manoeuvre.This enables it to perform a remarkable range of stunts including track stands, rear-wheel hops and jumps.While the current prototype is far too small to carry a person, its ability to balance and clear obstacles could one day inform the design of ultra-agile personal vehicles capable of travelling where conventional bikes and scooters cannot. At the heart of the design is a powerful moving mass that allows the robot to shift its weight and control its momentum in mid-manoeuvreThe UVM, which weighs 23.5kg, has been developed by a team at the Robotics and AI Institute based in Massachusetts.It can move at eight metres per second and jump over obstacles that are 130 per cent of the robot’s height.Writing in the journal Science Robotics, where the design was published, the researchers said: ‘Trail cyclists and mountain bike riders can hop, jump, balance, and drive on one or both wheels.‘This versatility allows them to achieve speed and energy-efficiency on smooth terrain and agility over rough terrain.‘Inspired by these athletes, we present the design and control of a robotic platform, Ultra Mobility Vehicle (UMV), that combines a bicycle and a reaction mass to move dynamically with minimal actuated degrees of freedom.’To create the machine, the team used computer simulations to test dozens of different designs before settling on the final bike-like layout.They specifically optimised the robot for two key abilities – jumping as high as possible and maintaining balance while riding on a single wheel. The AI-powered robot bicycle can pull wheelies, leap over obstacles and even perform front flipsRather than relying on kickstands, stabilisers or heavy reaction wheels, the UMV stays upright by constantly shifting its weight and using its momentum to correct its position. Most of the robot's mass is concentrated in a movable 'head' mounted above the bike frame, allowing it to throw its body around much like a human rider performing tricks.Engineers then trained the machine in a virtual environment using reinforcement learning, a form of artificial intelligence in which software learns through trial and error.The robot practised thousands of simulated manoeuvres before successfully transferring its skills to the real world.During testing, the UMV learned a range of behaviours including riding forwards and backwards, balancing stationary in a 'track stand', hopping sideways on a single wheel and turning on the spot.Researchers said one of its most surprising abilities was a 'shimmy turn', which emerged naturally from the AI system rather than being explicitly programmed. The device mimics the gravity-defying stunts performed on the Batpod in the Dark Knight films (pictured)The robot can also autonomously approach obstacles, jump onto a one-metre-high platform, ride across it and drop off the other side before repeating the sequence.In another demonstration, it completed a full front flip in less than half a second before landing and riding away.According to the researchers, future versions could be used for urban deliveries, reconnaissance missions or traversing rough terrain where conventional bicycles and wheeled robots would struggle.‘UMV’s combination of speed and small spatial footprint could help with applications such as urban courier service, or long-distance reconnaissance over rough terrain,’ the researchers concluded.