Xiaochan Xue | Yao Zheng@UHM

DOE Genesis Mission: STRATOS: Security and Trust Runtime Architecture for Time-critical Operational Science

Mon, 17 Aug 2026 10:10:00 -1000

Executive Summary

STRATOS develops a cloud-native, model-agnostic security middleware for protecting AI models used in time-critical scientific and energy-system operations. AI is increasingly embedded in workflows such as grid forecasting, contingency analysis, and operational decision making, creating new risks from adversarial perturbations, inference-time evasion, and backdoor attacks that may remain physically plausible while corrupting model outputs. STRATOS addresses this problem through a physics-grounded DevSecOps architecture that continuously evaluates inference streams, computes certified trust scores, and distinguishes malicious manipulation from legitimate operational variability without requiring modification of the protected model.

The project combines physics-constrained adversarial emulation, imbalance-resilient certified detection, and a multi-tier runtime mitigation pipeline that includes targeted input purification, Control Barrier Function-based graceful degradation, and quarantine or rollback of compromised data and models. The Phase I system will be evaluated across the University of Hawaiʻi at Mānoa campus microgrid and Argonne National Laboratory’s Controller-Hardware-in-the-Loop platform, targeting at least 90% certified detection, no more than 5% false positives, at least 95% interception of synthesized adversarial payloads, and end-to-end latency of 20 ms or less. The longer-term goal is to transition STRATOS toward a federated security architecture for trustworthy AI across DOE scientific computing and critical-infrastructure environments.

NVIDIA Academic Grant Program: OmniPort: Real-Time Semantic Digital Twins via ISAC and AI-RAN for Smart Ports

Sat, 15 Aug 2026 15:29:00 -1000

Executive Summary

OmniPort develops a real-time intelligent wireless and digital-twin platform for safer and more efficient autonomous operations in container ports. Ports are particularly challenging environments for robots because stacked metal containers obstruct GPS and line-of-sight sensors while producing strong wireless multipath. Instead of treating these reflections solely as interference, OmniPort uses integrated sensing and communication (ISAC) to extract information from routine robot uplink signals, enabling around-corner hazard detection and GPS-free robot localization. An AI-RAN then combines these sensing results with robot mobility and network telemetry to predict wireless conditions, dynamically allocate network resources, and maintain low-latency communication for safety-critical operations.

The project integrates these capabilities into an uncertainty-aware semantic digital twin that represents container geometry, robot movement, hazards, and wireless connectivity in real time. Rather than continuously transmitting camera video, operators can supervise robot fleets through a privacy-preserving VR environment generated from this semantic information, with warnings for collision risks, connectivity degradation, and emerging hazards. The platform uses four on-premises NVIDIA RTX PRO 6000 GPUs to support concurrent ISAC processing, AI-RAN control, NVIDIA Isaac Sim/Cosmos-based digital-twin reasoning, and Omniverse VR visualization, targeting an end-to-end control latency below 50 ms. The resulting technologies, software, and datasets are intended to provide a foundation for intelligent robotic operations in Hawaiʻi’s ports and other complex industrial environments.

Artificial Intelligence Radio Access Network (AI-RAN) with Digital Twin

Mon, 01 Sep 2025 04:14:54 -0800

Executive Summary

This lab-scale mmWave AI-Based RAN testbed integrates OAIBox, NVIDIA Aerial RAN, NI USRP X410, TMYTEK mmWave beamformers, and a high-fidelity Digital Twin pipeline using Remcom Wireless InSite and ANSYS HFSS SBR+ to create a flexible, programmable, and AI-native 5G/6G research environment. It enables real-time prototyping of mmWave physical layers, AI-driven beam management, hybrid beamforming, AI-enhanced MAC scheduling, and joint communication–sensing (ISAC) experiments. The platform supports end-to-end 5G NR PHY/MAC stacks, GPU-accelerated baseband processing, and mmWave RF front-ends for high-bandwidth OTA testing, while the Digital Twin provides physics-accurate ray-tracing, EM-based antenna modeling, and virtual–to–real co-simulation for channel prediction, beam optimization, and AI dataset generation.

Core Components

OAIBox – OpenAirInterface RAN Framework

The testbed uses OAIBox as a compact and modular implementation of the full OAI RAN stack. It provides:

Support for 5G SA/NSA gNB and UE
Flexible PHY–MAC integration
Customizable scheduling, HARQ, and protocol features
Real-time experimentation with RAN procedures and RRC signaling

OAIBox acts as the protocol and control anchor of the testbed.

NVIDIA Aerial RAN (cuPHY + cuMAC)

The NVIDIA Aerial platform provides GPU-accelerated baseband processing and AI-native PHY/MAC capabilities:

cuPHY for NR physical-layer DSP on GPUs
cuMAC for dynamic MAC scheduling on GPU
TensorRT for real-time neural inference
Support for multi-cell, multi-user, and high-throughput pipelines

Aerial enables experiments in:

AI-driven beam selection and prediction
Neural channel estimation
Predictive link adaptation and blockage detection

NI USRP X410 – Wideband Software-Defined Radio

The USRP X410 serves as the flexible transceiver frontend with:

Up to 400 MHz instantaneous bandwidth
Four synchronized TX/RX channels
10/1588 PTP synchronization
Digital IF for integration with NVIDIA Aerial

It supports:

mmWave IF/baseband experimentation
Real-time CSI acquisition
Multi-subarray MIMO and wideband waveform prototyping

TMYTEK mmWave Beamformers (BBox, UD-Box, Beamform Modules)

TMYTEK hardware provides programmable mmWave RF front-ends:

UD-Box for 24–32 GHz up/down-conversion
BBox One / BBox Lite beamforming arrays
API-driven phase/gain control
Rapid beam steering and codebook-based operation

These modules enable:

Hybrid or analog beamforming
Electronic steering up to ±60°
Multi-beam and multi-focus mmWave experimentation

Digital Twin

A full Digital Twin framework integrates high-fidelity electromagnetic simulation with the physical testbed. This enables reproducible channel modeling, data augmentation, and virtual-to-real RAN optimization.

Remcom Wireless InSite – Ray Tracing Propagation

Wireless InSite provides a large-scale propagation environment supporting:

GPU-accelerated 3D ray tracing
Detailed mmWave diffraction, reflection, and scattering
Urban, indoor, and open-field scenario modeling
Material-dependent loss and blockage effects
Beam-level channel prediction

ANSYS HFSS SBR+ – Full-Wave EM Modeling

HFSS SBR+ enables full-wave modeling of:

Antenna arrays, including TMYTEK beamformers
Reflectarrays, metasurfaces, and RIS
Realistic gain patterns for hybrid beamforming
Complex EM interactions under mmWave frequencies

Capabilities

AI-Enhanced RAN Intelligence

Neural beam prediction and tracking
AI-based MAC scheduling
CSI-driven link adaptation models
Blockage prediction and proactive beam switching

mmWave PHY/MAC Research

Hybrid and digital beamforming
Channel sounding and dataset generation
Evaluation of mobility, rotation, and blockage
5G NR waveform prototyping

Flexible RAN Architecture

O-RAN 7.2 split between OAIBox and Aerial
Multi-RU and multi-sector emulation
Edge-cloud cooperative intelligence

ISAC (Integrated Sensing and Communication) Extensions

Joint radar–communication waveform experiments
2D/3D angle estimation
Passive sensing with GPU-accelerated FFT pipelines

Digital Twin–Driven Insights

Predictive channel statistics and blockage maps
Virtual scenario pre-testing
Dataset augmentation for AI training
Virtual beam codebook optimization

RASRE: Reflectarray and Applications in Smart Radio Environment

Tue, 10 Dec 2024 00:00:00 -1000

Executive Summary

Reflectarray present a paradigm shift in wireless communication and sensing in mmWave and sub-THz region by offering unprecedented control over the propagation environment. Traditional wireless systems are largely at the mercy of the channel, which can be severely impaired by obstacles, fading, and interference. Reflectarray, however, introduce a new degree of freedom by enabling programmable manipulation of the wireless channel. Through precise control of the reflection and refraction properties of an array of elements, reflectarray can be used to mitigate path loss, combat fading, null interference, and shape radiation patterns, which can be used to enhancing spectral efficiency, energy efficiency, and expanding coverage. This project aims to investigate the fundamental principles of reflectarray operation, develop novel reflectarray architectures and control algorithms, and explore their applications in a wide range of wireless communication, sensing, energy transfer, and security scenarios.

Logistics

CRN

Semester ENGR196 ENGR296 ENGR396 ECE496

Spring 2025 TBD TBD TBD TBD
Personnel

Advisor Office Hours

Yao Zheng

Email with ‘’[VIP RASRE]’’ in the subject line.

See
Meeting:

Time Location

TBD HH488, Zoom
Workload

ENGR196 ENGR296 ENGR396 EE496¹ Deliverables²

$ \geq $ 3 H/W $ \geq $ 3 H/W $ \geq $ 6 H/W $ \geq $ 8 H/W

One per group

Indicate the authors of the appropriate sections.

Semester	ENGR196	ENGR296	ENGR396	ECE496
Spring 2025	TBD	TBD	TBD	TBD

Advisor	Office Hours
Yao Zheng Email with ‘’[VIP RASRE]’’ in the subject line.	See

Time	Location
TBD	HH488, Zoom

ENGR196	ENGR296	ENGR396	EE496¹	Deliverables²
$ \geq $ 3 H/W	$ \geq $ 3 H/W	$ \geq $ 6 H/W	$ \geq $ 8 H/W	One per group Indicate the authors of the appropriate sections.

Grading

General Breakdown

Category	Portion of Grade
1- Quality of project work Engagement, pursuit of knowledge necessary for project, contributions to technical progress of project, (396 and 496: contributions to management of project).	35%
2- Documentation and records Design notebook, wiki documentation, presentations, Reports.	35%
3- Teamwork and interaction Attendance to meetings, contributions to team, peer evaluations, team presentations and reports	30%

Breakdown by Deliverable

Deliverable	ENGR 196/296/396	EE496/499	Applicable Categories
Project proposal report	10%	10%	1,2,3
Preliminary design review presentation	15%	20%	1,2,3
Design notebook check	10%	5%	1,2
VIP poster session presentation	10%	15%	1,2,3
Final report draft (proofed by peers)	N/A	N/A	1,2,3
Final recorded video presentation	20%	25%	1,2,3
Final report	25%	15%	1,2,3
Design notebook final check	10%	10%	1,2

Cutoffs

A- B- C-

70% 50% 30%

A-	B-	C-
70%	50%	30%

Iterations

SEMESTER	TOPICS	DOCUMENTS
Spring, 2025	Designing 3.5GHz reflectarray with transistors

This course adheres to UH Manoa W focus requirements and ECE Department requirements. So you are required to write 4,000 words or more throughout the semester, split between the proposal and the final report. ↩︎
Reports should be uploaded to the Google Drive before the deadline. Late reports are subject to a minimum 15% reduction in grade. Presentation files (PowerPoint or PDF) should be uploaded to the Google Drive before or immediately after your presentation. ↩︎