C2S: An Transformer-Based Framework to Extrapolate Sensing Channel From Communication Channel

Next-generation mobile networks are poised to leverage Integrated Sensing and Communication (ISAC) as a pivotal technology, offering unprecedented support for various sectors including industrial Internet of Things (IIoT), extended reality (XR), and smart home applications. A critical challenge in implementing ISAC lies in extracting sensing parameters from radio signals, a task that has proven difficult using conventional methods. This paper, for the first time, introduces a novel approach to this challenge by proposing the extrapolation of sensing channel information, specifically the power delay profile (PDP), from readily available communication channel state information (CSI). We present a transformer-based framework designed to accomplish this task efficiently. Our extensive simulations demonstrate the framework's effectiveness in CSI-to-PDP extrapolation, achieving high accuracy in predicting both the delay and signal strength of individual paths. This innovative method has the potential to significantly enhance sensing capabilities in future mobile networks, paving the way for more robust and versatile ISAC applications.