Principal Investigator: Professor Chen Xudong, ECE

Millimeter wave (MMW) multiple input multiple output (MIMO) imaging technologies are widely used in various security applications. Real-time three-dimensional (3D) computational imaging becomes increasingly significant with the widespread deployment of 5G base stations, since MMW and massive-MIMO are key components of 5G technology. The 5G devices, components, and nodes are specifically designed for effective wireless communication, but we can make good use of the signals between pairs of nodes to image objects within network area as well. The proposed research aims to provide fast 3D MIMO MMW computational imaging via physics-inspired machine learning (ML) algorithms. Here, to avoid using ML as a purely data driven black-box solver, we emphasize on how to profitably combine ML with the domain knowledge on wave physics. Imaging at MMW spectrum is not the type of "what you see is what you get", and thus the wave physics that is involved in the imaging process is very important. The proposed physics-inspired machine learning can achieve fast 3D imaging for MMW MIMO system, and it can be used in a variety of real-world applications, such as surveillance in private or public regions, patient movement tracking, and monitoring traffic situation.