Zhejiang University researchers have taken a significant leap forward by creating an intelligent aeroamphibious invisibility cloak capable of operating seamlessly across sea, land, and air environments. This advancement, published in Advanced Photonics, introduces a self-operating cloaked drone that cleverly blends perception, decision-making, and execution through spatiotemporal modulation of reconfigurable metasurfaces.
For decades, scientists have faced the challenge of designing an invisibility cloak that can adapt in real time to dynamic landscapes, a task complicated by the need for advanced metasurfaces and algorithms to overcome issues like non-uniqueness and incomplete inputs. The team at Zhejiang University tackled these obstacles by developing a generation-elimination neural network or stochastic-evolution learning. This innovative approach allows the spatiotemporal metasurfaces to globally adapt, finding optimal configurations that achieve invisibility across varying terrains.
The breakthrough experiment demonstrated the drone's ability to achieve adaptive invisibility in three distinct environments, marking a significant advancement in invisibility technology. This integration of deep learning, advanced control systems, and spatiotemporal metasurfaces opens new doors for inverse design, solving complex many-to-many correspondences and pushing the boundaries of materials science and adaptive device development.
While this achievement represents a major step forward, the research also acknowledges existing limitations such as bandwidth constraints and the challenges of full polarization. Nonetheless, this work lays the groundwork for future advancements in the field, promising new directions for research in materials discovery and the creation of adaptive metadevices.
Research Report:Autonomous aeroamphibious invisibility cloak with stochastic-evolution learning
Related Links
International Society for Optics and Photonics
Space Technology News - Applications and Research
| Subscribe Free To Our Daily Newsletters |
| Subscribe Free To Our Daily Newsletters |
