Biography
Jui Mhatre is a PhD candidate and Graduate Research Assistant in Computer Science at Kennesaw State University. Her research focuses on computer networks, 5G/6G wireless communications, AI-driven networking, and IoT systems. Her current work explores the use of artificial intelligence and wireless channel intelligence to improve channel prediction, reduce signaling and pilot overhead, and enable more efficient next-generation wireless networks. She has also worked extensively on LoRaWAN-based low-power IoT systems and AI-driven network optimization. Prior to her doctoral studies, she worked as a software developer and gained industry experience in machine learning and software engineering.
Abstract
Title: AI-Driven Channel Intelligence for Next-Generation Wireless Networks
Abstract: As wireless networks evolve toward 5G-Advanced and 6G, the ability to acquire accurate channel state information is becoming increasingly important for high-frequency, wideband, and large-scale antenna systems. At the same time, conventional channel estimation and CSI feedback mechanisms introduce substantial pilot, signaling, and energy overhead. This raises an important question: can wireless networks predict their channel conditions using information they already have, rather than continuously requesting additional measurements?
This presentation explores that question through an AI-driven approach to downlink channel prediction. The proposed framework, Geometry-Aware Neural Radio Transformer (GA-NRT), combines uplink channel observations with information about the surrounding physical environment to predict downlink channel characteristics without requiring dedicated downlink pilots during inference.
The presentation will discuss how transformer-based learning, graph representations of wireless propagation environments, and cross-attention can be combined to capture relationships between radio-frequency observations and physical propagation geometry.
The approach is evaluated using realistic ray-tracing-based wireless channels across different frequency bands and propagation conditions. Results demonstrate the potential to accurately reconstruct downlink channel information while reducing reliance on conventional pilot signaling.
The presentation will conclude by discussing how AI, wireless sensing, and environmental intelligence can work together to build more efficient, adaptive, and intelligent 5G/6G networks.
Jui Mhatre is a PhD candidate and Graduate Research Assistant in Computer Science at Kennesaw State University. Her research focuses on computer networks, 5G/6G wireless communications, AI-driven networking, and low-power IoT systems, with particular interests in intelligent channel estimation and reducing communication and signaling overhead in next-generation wireless networks.
