科学101 | 智能电磁隐身
世间万物能像哈利波特小说中一样隐身吗？
动态目标识别能像孙悟空一样“火眼金睛”吗？
带着这两个问题，钱超研究员开启了基于类脑超材料的动态光学幻觉之旅。
作者简介
ABOUT THE AUTHOR
钱超博士
ZJUI助理教授、研究员
研究方向包括新型人工电磁材料、电磁隐身、机器学习、光计算等。目前已发表40余篇SCI论文，其中以第一/通讯作者在Nature Photonics、Nature Communications (4篇)、Science Advances、Physical Review Letters、Advanced Materials、Light: Science & Applications (2篇)等国际期刊上发表论文30篇，9篇ESI高被引论文，获Editors' Suggestion论文和最佳论文主编推荐奖；系列成果多次被Physicsworld、Physics、Phys.org、EurekAlert等科技杂志报道。授权国家发明专利10余项，主持参与国家自然科学基金青年项目，国家重点研发项目等。入选中国科协青年人才托举工程、嘉兴星耀南湖青年人才等。
QIAN Chao
ZJUI Assistant Professor
Chao’s research is on basic science and applications of metamaterials, machine learning, and optical computing. He, as the first author, has published over 40 papers in high-profile journals, including Nature Photonics, Physical Review Letters, and Light: Science & Applications. He now has over 10 patents and presided over and participated in the National Natural Science Foundation of China's youth projects, national key research and development projects. He has been selected as one of the young talents of the China Association for Science and Technology (CAST) and Jiaxing South Lake Young Talents Program.
Slide to English Version
科研成果介绍
INTRODUCTION
目标识别在计算机视觉和图像处理领域具有重要意义，但传统目标识别方法在快速信息收集、并行数据处理、低能耗计算方面存在挑战。此外，传统的隐身技术在电磁参数、设计制备等方面也存在着困难。针对上述问题，钱超团队利用超材料和人工智能技术手段的交叉研究，提供创新解决方案，实现高效的目标识别和智能隐身。
Target recognition plays an important role in the fields of computer vision and image processing, but traditional target recognition methods face challenges in rapid information acquisition, parallel data processing, and low-power computing. In addition, traditional stealth technology faces difficulties in electromagnetic parameters, design and fabrication. In response to these problems, our research aims to provide innovative solutions using the technologies of metamaterials and artificial intelligence to achieve efficient target recognition and intelligent stealth.
Slide to English Version
基于动态目标识别和幻觉隐身的类脑超材料
A brain-inspired metamaterial for dynamic object recognition and illusion cloaking
钱超团队通过将深度学习、拓扑优化算法与类脑超材料相结合，实现了动态三维物体的自动识别。类脑超材料具备被动和主动两种智能形式，能够对散射电磁波进行处理分析，并实时识别物体的姿势。此外，团队还提出了一种基于类脑超材料的动态光学幻觉方法（图1），能够将连续动作转换为不同的视觉效果，为光学信息处理和显示等领域提供了新的可能性，有望开启“光学万物互联”的新时代。
图1 基于类脑超材料的动态光学幻觉
类脑超材料可以应用于光安检系统，为安检系统提供高效的目标识别和分析能力（图2）。此外，它还可以应用于无线通信领域，提高通信系统的抗干扰和安全性能。同时，基于类脑超材料的动态光学幻觉方法也可以应用于光学信息处理和显示领域，创造出更加逼真和沉浸式的视觉效果。
图2 类脑超材料在光安检系统中的应用示意图
By combining deep learning and topology optimization algorithms with brain-inspired metamaterials, the team has achieved automatic recognition of dynamic 3D objects. Brain-inspired metamaterials possess passive and active forms of intelligence, enabling the analysis of scattered electromagnetic waves and real-time recognition of object poses. Furthermore, they have proposed a dynamic optical illusion method based on brain-inspired metamaterials (Figure 1), which can transform continuous actions into different visual effects, providing new possibilities for optical information processing and display, and potentially opening a new era of "optical connectivity for everything."
Figure 1: Dynamic optical illusion based on brain-inspired metamaterials
The brain-inspired metamaterial can be applied in optical security inspection systems, providing efficient target recognition and analysis capabilities, as shown in Figure 4. Additionally, it can be applied in the field of wireless communications to enhance the anti-interference and security performance of communication systems. Moreover, the dynamic optical illusion method based on brain-inspired metamaterials can be applied in the fields of optical information processing and display, creating more realistic and immersive visual effects.
Figure 2: Application of brain-inspired metamaterials in optical security inspection systems
Slide to English Version
基于深度学习的新一代智能隐身器件
A new generation of intelligent stealth devices based on deep learning
钱超团队通过建立时域电磁仿真模型、设计智能电磁隐身材料，并结合深度学习训练和实验数据采集，实现了智能隐身器件对动态电磁环境的实时响应和自适应隐身，成功实现了微波段的智能自适应隐身系统的搭建（图3）。
图3 智能隐身系统框架
团队为了测试智能隐身衣的性能，开展了大量实验，从近场成像和远场散射截面测量等多方面验证了智能隐身衣样机的可行性和鲁棒性（图4）。这项研究突破了传统隐身技术的限制，为实现更广泛的应用场景提供了新的可能性。
图4 智能隐身实验装置
We have successfully developed an intelligent adaptive stealth system for the microwave band (Figure 2). By establishing a time-domain electromagnetic simulation model, designing intelligent electromagnetic stealth materials, and combining deep learning training and experimental data collection, we have achieved real-time response and adaptive stealth of the intelligent stealth device to dynamic electromagnetic environments.
Figure 3: Framework of the intelligent stealth system
To test the performance of the intelligent stealth fabric, the team conducted a large number of experiments and verified the feasibility and robustness of the intelligent stealth fabric prototype from various aspects such as near-field imaging and far-field scattering cross-section measurements (Figure 3). This research breakthrough overcomes the limitations of traditional stealth technology and provides new possibilities for a wider range of application scenarios.
Figure 4: Experimental setup for intelligent stealth
Slide to English Version
钱超团队的研究成果在超材料和人工智能领域取得了重要突破，分别实现了动态目标识别和幻觉隐身的类脑超材料，以及微波段的智能自适应隐身器件。这些成果在光安检系统、无线通信、幻觉隐身和隐身技术等领域具有广泛的应用价值，将为相关领域的发展和创新提供新的可能性。
The team’s research achievements have made significant breakthroughs in the fields of metamaterials and artificial intelligence, realizing brain-inspired metamaterials for dynamic object recognition and illusion cloaking, as well as intelligent adaptive stealth devices for the microwave band. These achievements have broad application value in areas such as optical security inspection systems, wireless communications, illusion cloaking, and stealth technology, providing new possibilities for the development and innovation of related fields.
Slide to English Version
作者说
AUTHOR'S NOTE
用于动态目标识别和幻觉隐身的类脑超材料这个炫酷的想法诞生于两个问题的碰撞。世间万物能像哈利波特小说中一样隐身吗？动态目标识别能像孙悟空一样“火眼金睛”吗？
然而，这个想法面临着巨大的挑战：需要设计合适的超材料的空间结构，并且能实现自动分析和处理散射电磁波。整个幻觉隐身系统的搭建需要包括材料科学、电磁学、计算机科学和深度学习等领域的知识技能的融合，这就需要整个团队在多个学科领域进行交叉研究，积极交流，共同攻克难题。
在解决这些困难的过程中，钱超研究员不断探索新的思路和方法，并与国内外团队成员进行深入讨论。他们在一次次讨论中不断改进他们的研究方案，并进行了大量的实验和仿真，不断调整提升系统的性能，以实现动态目标识别和幻觉隐身的目标。经过长期不懈的努力，钱超研究员将其脑海中的奇思妙想转换成了现实。值得一提的是，在利用兔子做实验的过程中（图5），小兔子活蹦乱跳时常不在有效区域内，钱超研究员亲自对小兔子进行引导，在实验过程中多次被小兔子咬伤，但是仍然坚持进行高标准高质量的实验，不轻言放弃。
图5 实验中的小兔子
在这种坚持不懈，开拓求真的科学家精神指引下，在交叉团队的长期的努力下，他们取得了惊人的成果。类脑超材料的研究为快速目标识别提供了新的途径，而基于深度学习的智能隐身器件则为隐身技术带来了新的突破。这些成果不仅展示了团队成员之间的紧密合作，也展示了他们对交叉学科的深入理解和应用。
The fascinating idea of using brain-like metamaterials for dynamic object recognition and illusionary invisibility emerged from the collision of two questions. Can anything in the world be invisible like in the Harry Potter novels? Can dynamic object recognition be as perceptive as Sun Wukong's "fiery eyes"?
However, this idea faced significant challenges. Addressing the requirements of the idea necessitated designing appropriate spatial structures for metamaterials that could automatically analyze and process scattered electromagnetic waves. Building the entire illusionary invisibility system required the fusion of knowledge and skills from material science, electromagnetics, computer science, and deep learning. This called for cross-disciplinary research within the team, active communication, and collaborative efforts to overcome the obstacles.
In the process of solving these difficulties, Dr. Qian Chao constantly explored new ideas and methods while engaging in in-depth discussions with team members from both domestic and international backgrounds. Through successive discussions, they continuously refined their research plans, conducted numerous experiments and simulations, and made iterative adjustments to enhance the system's performance, aiming to achieve the goals of dynamic object recognition and illusionary invisibility. After long and relentless efforts, Dr. Qian Chao transformed his imaginative ideas into reality.
It is worth mentioning that during the experiments with rabbits, the lively rabbits often moved out of the effective area. Dr. Qian Chao personally guided the rabbits and was bitten multiple times during the process. However, he persevered and carried out high-standard, high-quality experiments without giving up due to the bites.
Figure 5: rabbit in the experiment
Guided by the spirit of perseverance, exploration, and the pursuit of truth, and through the long-term efforts of the interdisciplinary team, they achieved astonishing results. The research on brain-like metamaterials provided a new approach for rapid object recognition, while the deep learning-based intelligent invisibility devices brought breakthroughs to the field of invisibility technology. These achievements not only demonstrated the close collaboration among team members but also showcased their profound understanding and application of cross-disciplinary knowledge.
Slide to English Version
学术成果奖
Academic Achievement Award
国际校区于每年年底评选出本年度最值得表彰的学术成果，2022年度共评选出本系列研究等10项研究成果，该研究成果为获奖作品之一。
International Campus selects the most commendable academic achievements of the year by the end of each year. In 2022, a total of 10 research achievements including this series of research were selected, this is one of the achievement.
Slide to English Version
图文 | 钱超
编辑整理 | 薛倩 任悦 李真鸣
图文编辑 | 李亦楠
责任编辑 | 吴锋滨