About Me

  • I’m Xiaowen Tian, a fourth-year Ph.D. student in Electrical Engineering at NC State University.
  • I am working as a graduate research assistant at NC State under the guidance of Prof. Robert W. Heath Jr and Prof. Nuria Gonzalez-Prelcic.
  • My research interest is in signal processing techniques in wireless communication systems. I am currently working on near-field communications.

Research Projects

Variable Beamwidth Near Field Beamforming in mmWave Systems

Previous approaches to operate in the near field provide very focused beams, which can only illuminate a small region around the target user. Therefore, this near field beamfocusing configuration is sensitive to location inaccuracy. It also increases the initial access time. We proposed a beamforming technique that has variable beamwidth by mapping the RIS elements to a tunable spherical surface and tuning the size and center of the surface.

Machine Learning-Assisted Codebook Design for Channel Estimation

Minimum mean-square error (MMSE) filters are applied for channel estimation which requires second-order statistics. However, they are not easy to acquire in practice. We found that auto-correlation functions (ACFs) are related to MMSE filters and channel parameters. Thus, with the assumption that similar channels can use one representative codeword for channel estimation, we proposed to design such a codebook to reduce the complexity during channel estimation.

Reconfigurable Intelligent Surface Configuration and Deployment under Near Field Model

With sub-THz frequencies, the links are easy to be blocked. The reconfigurable intelligent surfaces (RISs) can create additional non-line-of-sight (NLoS) links to overcome the blockage issue. However, large scale RISs are needed to meet the link budget, which makes the RISs electrically large and working in the near field mode. Far field configuration will cause a severe gain loss in this scenario, so we propose to configure and deploy the RIS smartly under the near field model.

Secure Beamforming in Millimeter Wave Systems

Millimeter wave channels are sparse, but Alice, Bob, and Eve may share some common scatters, which makes eavesdropping easier. However, such a special structure in the channel can also be exploited to null the signals transmitted through those common scatters to avoid eavesdropping.

Pilot Spoofing Detection and Secure Transmission

When Eve sends the same pilots as Bob during uplink transmission, Alice can be spoofed to transmit confidential signals to Eve during downlink transmission. In this project, we proposed a Random Training-Assisted (RTA) algorithm, which exploits the randomness of the pilots to detect the existence of Eve. In addition, we proposed enhancement transmission strategies for both detection results of present/absent Eve.