Generative AI for Integrated Sensing and Communication: Insights from the Physical Layer Perspective

Abstract

As generative artificial intelligence (GAl) models continue to evolve, their generative capabilities are increasingly enhanced, and being used exten-sively in content generation. Furthermore, GAl also excels in data modeling and analysis, benefiting wireless communication systems. In this article, we investigate applications of GAI in the physical layer and analyze its support for integrated sensing and communications (ISAC) systems. Specifically, we first provide an overview of GAI and ISAC, touching on GAl's potential support across multi-ple layers of ISAC. We then thoroughly investigate GAl's applications in the physical layer, such as channel estimation, which demonstrates the benefits that GAl-enhanced physical layer technologies bring to ISAC systems. Finally, in the case study, we present a diffusion model-based method for estimating signal direction of arrival in near-field scenarios using uniform linear arrays with antenna spacing over half the wavelength. With a mean square error of 1.03 degrees, the method confirms GAl's support for the physical layer in near-field sensing and communications.