NSF SWIFT-SAT / INTERACT — Learning-Based RFI Mitigation for Active–Passive Spectrum Coexistence (2024–Present)
- Designed a deep-reinforcement-learning (DDQN) spectrum-selection framework identifying interference-dominant 5G NR subbands for passive radiometer receivers operating in shared spectrum.
- Developed neural successive-interference-cancellation modules — U-Net time–frequency masking and a two-stage MLP — to suppress structured 5G OFDM interference while preserving passive radiometric baselines.
- Built end-to-end MATLAB/Python simulation and evaluation workflows (3GPP-compliant 5G NR waveform generation, Sionna CDL channel modeling, model training and baseline comparison).
- First-authored the IEEE DySPAN 2025 paper on AI-assisted successive interference cancellation and two follow-on manuscripts extending the framework.
NSF Convergence Accelerator Track G — Combating Vulnerability and Unawareness in 5G Network Security (2023–Present)
- First-authored the IEEE MILCOM 2024 paper on defending 5G networks against jamming attacks with multipath communications.
- Contributed to a three-path 5G Standalone architecture combining threshold secret sharing with path diversity for resilience against eavesdropping, jamming, and infrastructure failure.
- Contributed to an error-pattern steganographic communication framework concealing covert data within standard 5G traffic, and to hardware-in-the-loop validation on a 5G Standalone COTS testbed (srsRAN, Open5GS, USRP B210).
O-RAN, Federated Learning, and 6G Network Management (Dissertation Research, 2021–Present)
- Developed SliceFed, a federated constrained multi-agent deep-reinforcement-learning framework (Lagrangian primal–dual PPO with federated averaging) for dynamic spectrum slicing with URLLC latency guarantees in dense 6G radio access networks (submitted to IEEE DySPAN 2026).
- Evaluated reinforcement-learning schemes for trajectory optimization of aerial radio units in non-terrestrial 6G deployments (IEEE ICC 2023).
- Co-authored studies on AI-driven fuzzing for vulnerability assessment of O-RAN traffic-steering algorithms and on auditable geographic fairness for multi-operator LEO spectrum sharing.
AI-Enabled Radio Transceivers and Nonlinear Receivers
- Developed AI-driven demodulation approaches for nonlinear receivers operating in shared spectrum with high-power blockers (IEEE WCNC 2022).
- Co-authored a symbol-error-rate characterization of communication receivers under receiver nonlinearity (IEEE VTC-Spring 2020).
- Investigated opportunities and challenges of artificial-neural-network architectures for radio transceivers (IEEE VTC-Fall 2021).