Research

Next-generation Wireless Networks and Systems 

Next-generation wireless networks will support Gbps+ data rates and sub-millisecond latency, which can enable a broad range of real-time applications including AR/VR, autonomous driving, and smart cities. Bringing future wireless networks to reality requires significant research across all layers of the network stack. Our research focuses on both the theoretical and experimental aspects of a wide range of enabling technologies including millimeter-wave communicationsmassive-antenna systemsoptical-wireless communicationsedge cloud and computing, and full-duplex wireless. We also design practical, efficient, and scalable algorithms and systems, and develop customized prototypes and testbeds (such as the NSF PAWR COSMOS testbed) to evaluate their performance in real-world scenarios.

AI/ML-enabled Wireless and Optical Networking

As we are embracing 5G and beyond-5G mobile networks, their network infrastructures are in need of a revolution in order to fulfill the massive computation needs, increasing hardware and software complexities, and the large volume of cross-domain data. By leveraging the ongoing revolution in Artificial Intelligence (AI) and machine learning (ML), our research focuses on AI- and ML-powered wireless networking at the edge, with the goal to create a new adaptable, scalable, and performance-aware mobile network infrastructure that can provide flexible services for heterogeneous use cases. We will also explore a data-driven approach to complement and augment existing algorithmic alternatives in the design of next-generation wireless networks and systems.

Spectrum Monitoring and Coexistence

The expanded spectrum usage in the 5G and beyond-5G eras inevitably calls for the coexistence of commercial services (e.g., cellular networks), non-commercial active users (e.g., weather satellite and GPS), and passive users (e.g., radio astronomy), leading to interference for both active and passive users. Our research focuses on the design of a cooperative network system for spectrum coexistence, where each receiver (RX) can detect interference in real-time, identify the type and source of interference, and more importantly, react to the interference by adopting selective interferer nulling, leveraging an intelligent network control architecture and machine learning (ML) techniques.

Acknowledgments: Our research projects are supported in part by NSF grants CNS-2128638 (NSF SWIFT) and CNS-2112562 (NSF NAI), as well as a Google Research Scholar Award, an IBM Academic Award, and an ACM SIGMOBILE Student Community Grant.