Research Projects

Massive MIMO is a cornerstone of modern cellular systems, but realizing its full potential depends on efficient beam management and channel state information (CSI) acquisition. Our research develops AI-enhanced beam management techniques that improve the efficiency of 5G NR and future 6G MIMO systems while remaining fully compatible with existing 3GPP standards. Rather than replacing standardized signaling procedures, we investigate how machine learning can intelligently augment them to reduce training overhead, improve beam selection, and enhance CSI quality. 

A major focus of our work is the design of AI-enabled 3GPP-compliant codebooks and learning-assisted beam training for Type-I and Type-II CSI feedback. We investigate learned beam codebooks tailored to realistic propagation environments, AI-based CSI feedback generation that bypasses conventional channel estimation and exhaustive beam search, and low-overhead beam management techniques for massive MIMO systems. Our long-term vision is to develop AI-native beam management frameworks that preserve interoperability with 3GPP standards while enabling scalable, low-complexity, and high-performance massive MIMO operation for future wireless networks. 

Future 6G networks are expected to expand into the upper mid-band (FR3, 7-24 GHz), where large portions of the spectrum are already occupied by incumbent satellite services. Our research develops signal processing, optimization, and AI-driven resource management techniques that enable terrestrial cellular and satellite networks to efficiently share FR3 spectrum while maximizing terrestrial performance and protecting incumbent satellite users.

Our current work focuses on coexistence-aware beamforming, where cellular transmissions are jointly optimized to maximize network throughput while satisfying explicit interference protection requirements for incumbent satellite terminals. Looking ahead, our research will expand in two complementary directions. The first investigates dynamic spectrum coexistence between independent terrestrial and satellite systems through beam tracking for moving LEO satellites and mobile satellite terminals, AI-native spectrum sharing and predictive interference management. The second explores integrated terrestrial-non-terrestrial networks (TN-NTNs), where terrestrial and satellite infrastructures cooperate through joint beamforming, resource allocation, and mobility management to provide seamless global connectivity. Our long-term vision is to develop intelligent algorithms that enable both efficient spectrum sharing and tightly integrated satellite-terrestrial networks for 6G and beyond. 

The emergence of the low-altitude economy is driving the need for ubiquitous, reliable, and scalable wireless connectivity for UAVs. Our research investigates how existing 4G and 5G cellular networks can be intelligently evolved to support aerial users while preserving the quality of service of terrestrial users. We develop sustainable and cost-effective solutions that maximize the reuse of existing cellular infrastructure through intelligent architecture design, advanced beamforming, and radio resource management. 

A major focus of our work is the design of hybrid cellular architectures, where dedicated uptilted antennas are collocated on selected terrestrial base stations to provide ubiquitous aerial coverage without compromising terrestrial users performance. We investigate beam management, cell association, interference management, and radio resource allocation for aerial users, including extensions of 5G NR beam management procedures to support UAV mobility. Our long-term vision is to develop intelligent cellular networks that seamlessly support both terrestrial and aerial users as an integral part of future 6G ecosystems. 

Integrated Access and Backhaul (IAB) is emerging as a key technology for extending network coverage and densifying future wireless systems without requiring fiber connectivity at every base station. Our research investigates next-generation IAB architectures that intelligently exploit wireless relays to provide seamless connectivity while jointly optimizing access and backhaul performance. We develop signal processing and radio resource management techniques that maximize end-to-end network throughput while accounting for the inherent coupling between access and backhaul links.

Our work spans multiple relay architectures, including terrestrial small cells, UAV-assisted relays, and distributed IAB deployments. A major focus is understanding which relay architecture is best suited for different deployment scenarios and how relay placement, user association, beamforming, and radio resource allocation should be jointly optimized based on the propagation characteristics of both access and backhaul links. Looking ahead, we are extending this research toward distributed and cell-free MIMO IAB, cooperative multi-hop relaying, and AI-driven network topology optimization for scalable 6G deployments.