Yuzhong Wang, Yuxuan Wang, Ziang Yue, Yuxuan Wang, Xinyuan Xu, Yanfang Li, Qiming Wang, Cheng Pang, Yifei Wang, Pengcheng Wang, Yongkang Dong, Yayun Cheng, Ari Sihvola, Jiaran Qi
The critical role of personnel security screening (PSS) devices in modern public transportation has driven a growing demand for real-time, lightweight, noninteractive, and high-performance screening systems. Currently, the predominant methods involve handheld metal scanners that require active cooperation by the subject, generally suffering from low efficiency, limited detection range, and blind spots for nonmetallic objects. Millimeter-wave-based PSS technology, featured by its speediness, harmlessness, and capability of detecting nonmetallic materials, represents an emerging trend in security screening. However, traditional millimeter-wave system-wide dynamic scanning schemes face challenges such as bulky size, high complexity, and poor integration. To address these problems, we develop a wide-angle metalens based on a neural-network-assisted end-to-end inverse design method, aiming to replace thick and heavy dielectric lenses in traditional millimeter-wave PSS systems, thereby enhancing system integration and lightweight capabilities while improving wide-angle scanning performance. As proof of concept, we propose a system-wide static noninteractive wide-angle dual-polarized meta-detector, which consists of a dynamic beam steering transmitter, dual-polarized receivers, and a data acquisition module. This meta-detector can rapidly, precisely, and statically identify concealed objects of various materials on the human body in a noninteractive manner within a 32° horizontal and 26.8° vertical field-of-view at an operating range of 15–25 m. The proposed PSS system offers long detection range, high integration, and noninteractive operation, highlighting its potential for practical deployment in real-world public security screening scenarios.