Still in experimental stages, the technology could prove vital in future applications such as search-and-rescue missions or even exploration inside complex cave systems
Pohang University
Engineers have developed a flying squirrel-inspired drone capable of agile movement in tight spaces by mimicking nature and using machine learning. Unlike traditional quadcopters, which struggle with limited manoeuvrability due to unidirectional thrust, this drone features foldable silicone wings that adapt mid-flight.
Controlled by reinforcement learning trained on human flight demonstrations, the system adjusts wing positions to optimise drag and lift, even compensating when mechanical thrust is maxed out.
Still in experimental stages, the technology could prove vital in future applications such as search-and-rescue missions or even exploration inside complex cave systems.
