Agile asymmetric multi-legged locomotion: contact planning via geometric mechanics and spin model duality
This bizarre new gait makes robots faster by breaking all the rules of symmetry.
Deep Dive
MIT researchers have developed a new AI control framework for multi-legged robots that breaks traditional symmetry. By applying geometric mechanics and a 'spin model duality' from statistical physics, they discovered an asymmetric gait for a hexapod robot. This new strategy achieves a forward speed of 0.61 body lengths per cycle, a 50% improvement over conventional gaits. The method even allows two legs to be completely unactuated and replaced with rigid parts.
Why It Matters
This could lead to radically more efficient and simpler-to-build robots for logistics and exploration.