Seeing the Spectrum: GPU-Accelerated ISAC for Secure, 5G-Enabled Spectrum Agility and Dynamic Coexistence

As wireless spectrum grows increasingly congested, efficient, and dynamic sharing is essential for meeting the requirements of systems that must share or coexist within the same band. In this talk, we explore how Integrated Sensing and Communication (ISAC), a capability poised to become a key feature in 6G and beyond, can be leveraged by commercial cellular systems to perform accurate spectrum sensing and reliably detect the presence of incumbent and primary spectrum users. We then examine how deploying edge AI on GPU-based platforms can significantly improve network responsiveness in performing real-time spectrum sensing. Building on this, we demonstrate how softwarized and open networks can dynamically adapt their operation to minimize performance degradation for all coexisting systems. Finally, we highlight the security implications of such an architecture and discuss the associated privacy considerations that arise from pervasive sensing.