Research & Publication

Research Scope
Integration of the state-of-the-art machine/deep learning (ML/DL) algorithms combined with signal processing techniques into wireless communications serves as the bedrock for charting pioneering directions in advancing Next Generation (NextG) wireless communication systems. Hence, our research agenda is directed to carve pathways towards the effective and efficient implementation of ML/DL in the evolutionary trajectory of NextG wireless with applications in 5G Ultra Wideband (5G-UW) ,  6G & beyond (6G&B) and terahertz (THz) communications, i.e., mmWave, especially in physical layer (PHY).

The envisioned progression towards enhancing performance of NextG wireless communication systems requires improvements in the accuracy and efficiency of PHY tasks. For example, this comes to significant importance when there is no definite operating models for channel modeling, estimation and signal equalization in THz communications. Hence, we follow following objectives to integrate ML/DL model into PHY operations:

1. To develop novel ML/DL architectures for various physical layer tasks, such as channel estimation, equalization, detection, decoding, modulation recognition, beamforming, and precoding.

   • This objective aims to address the challenges and limitations of conventional physical layer techniques, such as high complexity, low adaptability, and poor robustness when dealing with the dynamic and heterogeneous wireless environments.

   • ML/DL-based implementation of PHY tasks can offer advantages such as end-to-end optimization and nonlinear approximation which can improve the overall performance of physical layer tasks.

2. To investigate the integration of ML/DL with advanced wireless technologies, such as massive MIMO, THz, NOMA, and RIS.

   • This objective aims to explore the potential and challenges of applying ML/DL to enhance the performance of emerging wireless technologies that are expected to play a key role in 5G-UW and 6G&B.

   • ML/DL can help to overcome the difficulties and limitations in this domain such as high hardware cost, low spectral efficiency, severe propagation loss compensation and complex signal processing.

3.  To leverage the implementation of ML/DL in joint communication and sensing capabilities of THz waves for enabling services in 6G&B such as high-resolution imaging, localization, gesture recognition and biometric authentication.

   • This objective aims to exploit the unique features and opportunities of THz waves, such as ultra-high bandwidth, ultra-high resolution, ultra-low latency, and ultra-sensitive sensing.

   • ML/DL can help to extract deep correlative features from THz signals that enables new functionalities for 6G&B mobile network operations. 

4. To design ML/DL-based PHY security schemes for protection against various attacks such as jamming, eavesdropping, spoofing, and replay. 

   • This objective aims to enhance the security and privacy of wireless communication systems at the physical layer level, which is often neglected or compromised by traditional cryptographic methods.

   • ML/DL can help to detect and defend against various physical layer attacks by exploiting the characteristics of wireless channels or signals.

Publications

Since Google Scholar provides a comprehensive record of publications, SMARTEN lab's members' publications list can be accessed accordingly.

Faculty:    
Pejman Ghasemzadeh: Google Scholar