Mojtaba Vaezi
Home
Research
Group
Publications
Teaching
Service
Awards

Research

Text and Image Layout

My research interests lie at the intersection of communications, signal processing, machine learning, and network information theory, with a particular focus on sixth-generation (6G) radio access technologies, cyber-physical security, and medical imaging. I am dedicated to exploring the mathematical limits of sensing, storage, and communication and leveraging these insights to advance the design of smart sensing and communication networks. A key aspect of my recent work involves employing mathematical and learning-based tools to uncover these fundamental limits and developing practical methods to implement them in real-world scenarios.

We test our algorithms rigorously using software-defined radios (SDR) in the WIN Labs as well as on large-scale NSF platforms, ensuring practicality of solutions.

A descriptive image

Selected Projects
I have worked on several projects on large-scale wireless systems (including mobile and sensor networks) in collaboration with leading companies to design/optimize real-world complex systems. Some of my research projects, together with a statement of their potential impacts, are outlined below.

CAREER: Harnessing Interference with Deep Learning: Algorithms and Large-Scale Experiments

National Science Foundation, 2023-2028
 
    Inter-cell interference is a major challenge in achieving high throughput and low bit error rates in modern cellular networks, especially with the rise of dense, three-dimensional networks driven by drone adoption. Current solutions are limited by high signaling overhead and synchronization demands, often resorting to inefficient methods like resource orthogonalization or treating interference as noise. This project explores a novel approach using deep reinforcement learning to manage interference dynamically without requiring full channel information. Key objectives include designing interference-resilient networks, reducing bit error rates, and validating algorithms on large-scale experimental platforms.

    The implementation of this research involves developing and testing the proposed algorithms on the NSF POWDER and AERPAW platforms, which are large-scale experimental testbeds for advanced wireless research. The project will simulate real-world network environments, ensuring the algorithms are robust and scalable. By integrating cutting-edge machine learning models with practical system design, this research bridges the gap between theory and application, paving the way for intelligent interference management in real-life cellular networks.


Collaborative Research: NSF-AoF: AI-assisted Waveform and Beamforming Design for Integrated Sensing and Communication

National Science Foundation, 2024-2027
 
    This project focuses on developing AI-driven solutions for integrating sensing and communication functions to address challenges like limited spectrum and high hardware costs. By leveraging unified waveforms, beamforming designs, and large aperture arrays, the research aims to achieve significant gains in energy efficiency, cost-effectiveness, and system performance. The project employs multi-objective optimization and deep learning to tackle the trade-offs between sensing and communication metrics, enabling efficient and adaptive designs. The research includes three key thrusts:

  • Unified Design: Develop waveforms, constellations, and beamforming methods tailored for ISAC, along with AI-enabled channel learning.

  • Large Aperture Arrays: Explore the unique characteristics of extended near-field ISAC channels and optimize performance with electrically large aperture arrays (ELAA).

  • Real-Time Evaluation: Implement and evaluate AI-assisted ISAC designs for performance, robustness, and scalability.

    • This collaborative effort with Southern Illinois University, IL, and Aalto University, Finland, combines expertise from all three institutions to push the boundaries of ISAC research and its practical applications.


ERI: Interference-Aware Constellation Design for NOMA

National Science Foundation, 2023-2025


SDR-based implementation of NOMA


    The project focuses on a novel learning-based digital modulation design to mitigate inter-user interference in non-orthogonal multiple access (NOMA) systems. Specifically, it explores innovative superimposed constellations for NOMA, dynamically adapting to varying interference levels to reduce bit-error rates and minimize communication latency. The findings are expected to significantly impact both academic and industrial communities, providing opportunities for undergraduate involvement in deep learning for communications and contributing to the development of graduate-level coursework.

    Interference cancellation is a critical component of non-orthogonal systems. In this experimental work, we implemented a downlink, two-user NOMA system, where the user with stronger channel gain employs successive interference cancellation (SIC). This work aims to develop SIC-free decoding, as an alternative. The system was deployed on an NI USRP-2974 wireless testbed to evaluate its practical spectral efficiency. Experimental results in a real-world building environment demonstrate that NOMA improves spectral efficiency by approximately 1 bit/s/Hz compared to orthogonal multiple access (OMA).

Transmission Strategies for Sensor Networks

xyz
    Sensor networks require energy-limited communication/computing schemes, as sensors usually run on batteries, and it is required that the battery lasts for a long time. Distributed source coding (DSC) is a key enabling technology to this end. It refers to the compression of multiple, statistically dependent sources (e.g., neighboring sensors outputs) that do not communicate with each other and thus are encoded in a distributed manner. DSC has been privileged by the advancement of capacity-approaching channel codes, e.g., turbo and low-density parity-check (LDPC) codes. However, the performance of turbo/LDPC-based DSC systems deteriorates largely when the code length reduces. In practice, where delay and complexity limitations are stringent, code length cannot be very large. Thus, other low-delay and energy-limited communication strategies are crucial for sensor networks




Sincere Thanks to Our Sponsors