Khalifa University launches RF-GPT, an AI model that reads wireless signals
Khalifa University of Science and Technology announced RF-GPT, an AI language model designed to interpret radio-frequency signals directly. The model converts wireless signals into visual patterns that AI systems can analyze and describe in plain language, addressing a gap where existing language models operate only on text and structured network data.
The model outperformed baseline approaches by up to 75.4% on radio-frequency spectrogram tasks. It identified the number of signals in a spectrogram roughly 98% of the time-a task where general-purpose AI models rarely succeed.
How RF-GPT works
RF-GPT translates radio signals into spectrograms-visual representations of frequency data over time. The model then analyzes these patterns to answer questions about wireless spectrum activity using natural language.
Researchers trained the system on approximately 625,000 computer-generated radio signal examples. It handles tasks including signal identification, detecting overlapping transmissions, recognizing wireless standards, estimating device usage in Wi-Fi networks, and extracting data from 5G signals.
Intended users and applications
The model targets telecom operators, network engineering teams, and spectrum authorities managing increasingly complex wireless environments. Professor Merouane Debbah, Senior Director of Khalifa University's Digital Future Institute, said the work enables "a unified RF-language interface" where the physical layer becomes queryable in natural language.
Debbah framed the development as a step toward "AI-native 6G networks" where radio-frequency perception can directly support network optimization and policy decisions.
Research team and alignment
The project was led by researchers from Khalifa University, Université de Lorraine, and Zhejiang University. Professor Ahmed Al Durrah, Associate Provost for Research at Khalifa University, said the model reflects the institution's focus on AI integration across strategic sectors and aligns with the UAE National Artificial Intelligence Strategy.
The Digital Future Institute, which developed RF-GPT, combines foundational research with industry partnerships to build sector-specific foundation models and deployable AI platforms.
Your membership also unlocks:
