The Impact of beamforming in ISAC: a deep learning approach

Abstract

Integrated Sensing and Communication (ISAC) has emerged as a transformative paradigm in modern wireless networks, facilitating the seamless integration of sensing and communication functionalities within a shared spectrum and hardware framework. Unlike conventional frequency-division ISAC (FDSAC) techniques that allocate dedicated resources separately for sensing and communication, ISAC aims to enhance spectral and energy efficiency through advanced signal processing algorithms. The same beamforming vector is used for both communication and extracting information from targets [1]. Beamforming plays a crucial role in ISAC, significantly impacting both the communication rate (CR) and the sensing rate (SR). However, achieving an optimal trade-off between these two competing objectives presents a formidable challenge. Recent advances in Machine Learning (ML) have shown great potential in addressing complex optimization problems in wireless communications [2], [3], motivating us to compare the Deep Learning (DL)-based approach with the analytical approach to evaluate how DL can solve analytically complex problems in real time within communication systems.