A Survey on Machine Learning Enhanced Integrated Sensing and Communication Systems: Architectures, Algorithms, and Applications

Integrated sensing and communications (ISAC) technology is being developed in wireless communications systems in the sixth generation (6G). ISAC has the advantages of lower cost, better spectral efficiency, and better energy efficiency than systems that use separate transceivers and receivers. This...

Full description

Saved in:
Bibliographic Details
Published in:IEEE access 2024, Vol.12, p.170946-170964
Main Authors: Ade Krisna Respati, Mikael, Lee, Byung Moo
Format: Article
Language:English
Subjects:
5G mobile communication
6G mobile communication
activity recognition
Artificial intelligence
beamforming
channel estimation
deep learning
gesture recognition
Hardware
Integrated sensing and communication
Integrated signal and communication
localization
Machine learning
Radar
Receivers
signal processing
Surveys
Transmitters
waveform design
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Integrated sensing and communications (ISAC) technology is being developed in wireless communications systems in the sixth generation (6G). ISAC has the advantages of lower cost, better spectral efficiency, and better energy efficiency than systems that use separate transceivers and receivers. This technology utilizes the same communication resources for communicating and sensing within the same framework, enabling more efficient use of resources. Currently, machine learning (ML) has been developed in the field of communications, including sensing and wireless communications, due to its ability to tackle complex optimization problems, estimate complex issues, and extract and exploit spatial/temporal patterns that can improve ISAC performance. This paper provides a comprehensive survey of ISAC systems enhanced by ML. We begin by presenting various system configurations based on the type of radar and target sensing and the sensing source utilized in the ISAC system, as well as real-world ISAC use cases. Following an overview of ML and deep learning (DL), we explore common types of ML and DL models and their potential to enhance ISAC systems. We review the application of ML in ISAC systems to enhanced sensing performance and optimize ISAC signals. Finally, we outline the potential avenues for future research aiming to improve ML application on ISAC systems and other prospective applications.
ISSN:2169-3536
2169-3536
DOI:10.1109/ACCESS.2024.3501363