Ngo Van Mao

Projects

Mobile World Congress (MWC)

I led and delivered multiple AI-RAN and Open RAN system demonstrations at MWC Barcelona (2024–2026), with global industry partners, showcasing deployable AI-driven 5G/6G technologies.

MWC 2026

AI-RAN and O-RAN demos: YouTube playlist

  • Demo 1: Adaptive & Privacy-preserving Edge AI Model Partitioning Solution via Spectrum Sensing (SUTD, NeuroRAN, Keysight Technologies, LITEON)

    This demo considers optimizing AI model partitioning via spectrum sensing for split inference on visual analytics. Extending our MWC ‘25 demo, we will evaluate more realistic deployment scenarios with Vision Transformer (ViT) and demonstrate how to effectively balance energy efficiency and privacy preservation.

  • Demo 2: AI-on-RAN Orchestration: Enabling Real-Time Multimodal Intelligence for Autonomous Systems (SUTD, Red Hat, NeuroRAN, QCT)

    This demo showcases an AI-on-RAN Orchestrator that dynamically allocates and optimizes AI-RAN computing and communication resources to deliver efficient, differentiated AI services, enabling a real-time autonomous navigation robot powered by edge-based LLM inference.

    It further illustrates how an AI-native RAN can transform the network from a pure connectivity pipe into a revenue-generating platform for mobile network operators, supporting real-time, service-differentiated AI applications at the network edge.

  • Demo 3: QCT OmniRAN 5G O-RAN: Enabling AI-driven Energy-Saving rApp and Dynamic Network Slicing rApp (SUTD & QCT), and E-Plane Test Suite (Keysight Technologies)

    Co-developed with SUTD and validated by Keysight, this demo highlights how QCT can differentiate OmniRAN in the private 5G market by bundling deployable energy-saving and network slicing rApps that deliver measurable performance gains—including approximately 10% power reduction, while ensuring QoS assurance and supporting dynamic slice quota control.

  • Demo 4: Energy-Efficient O-RAN 5G Small Cells with Intelligent Energy-Saving rApp: LITEON-SUTD Collaboration

    LITEON and SUTD co-developed an energy-efficient O-RAN solution for 5G small cells, including an Energy-Saving rApp (ES-rApp) and a Traffic-Steering rApp (TS-rApp). SUTD contributed the ML-based predictive and energy-saving algorithms, while LITEON provided the 5G All-in-One platform and LiteNetics SMO. The ES-rApp achieved about 14% energy savings for capacity cells versus the baseline.

MWC 2025

AI-RAN and O-RAN demos: YouTube playlist

  • Demo 1: AI-Driven Spectrum Sensing for Dynamic & Privacy-Preserving AI Model Partitioning over 5G Network

    Together with Keysight, LiteON, and NeuroRAN, SUTD has developed an intelligent spectrum sensing system using the NVIDIA Aerial 5G Stack. The AI-driven spectrum sensing enables adaptive AI-driven ML model partitioning for privacy-focused image processing over 5G networks, addressing the inefficiencies of fixed model partitioning in dynamic wireless environments.

  • Demo 2: AI-Based 5G Beamforming for Mobility-Aware Interference Mitigation and Power Saving

    SUTD, in collaboration with Yonsei University and VIAVI, is showcasing the advanced deep reinforcement learning (DRL) for mobility-aware beamforming optimiser that addresses the interference issue and save energy when multiple massive MIMO base stations are deployed together in an urban scenario.

  • Demo 3: Accelerating O-RAN Adoption with Generative AI-Powered Network Operations

    In partnership with Amazon Web Services (AWS), StarHub, and Synergy Design Technology, SUTD has integrated Generative AI (GenAI) into O-RAN 5G networks, enabling telecom operators to monitor network performance, resolve issues, and optimise settings using GenAI-powered Agentic Assistant.

  • Demo 4: AI/ML and Digital Twin Enhanced O-RAN Energy-Saving Demonstration

    This demo, a collaboration between SUTD, VIAVI and Quanta Cloud Technology (QCT), showcases how Digital Twin technology can simulate and optimise energy-saving strategies in O-RAN-based 5G networks, thereby reducing power consumption without affecting performance.

  • Demo 5: O-RAN-Based Enterprise 5G Solution: From Lab to Field Trial

    Bridging Lab Testing and Real-World Deployment

    SUTD, together with StarHub, Synergy Design Technology, and VIAVI, demonstrates how O-RAN 5G technology transitions seamlessly from controlled lab environments to real-world enterprise applications, using 5G-powered drones and security robots.

  • Demo 6: Energy Efficiency Testing of Viettel High Tech Massive MIMO O-RU

    Advancing Energy Efficiency in O-RAN 5G Radio Unit

    In collaboration with Viettel High Tech (VHT) and Keysight, SUTD showcases the energy efficiency of VHT’s massive MIMO Open Radio Unit (O-RU), ensuring compliance with O-RAN standards while reducing power consumption.

  • Demo 7: Sustainable Global Orchestration for Service-Oriented Integration of Terrestrial and Non-Terrestrial Networks (TN-NTN)

    Terrestrial Network (TN) and Non-Terrestrial Networks (NTN)

    Partnering Japan’s National Institute of Information and Communications (NICT), SUTD presents a global orchestrator that dynamically manages terrestrial (TN) and satellite-based (NTN) networks to optimise energy efficiency and connectivity in remote regions.

MWC 2024

O-RAN system demos: YouTube playlist

  • Demo 1: Security Tests of QCT OmniRAN’s Handover Procedures by Asia & Pacific OTIC in Singapore (SUTD) using Keysight Test Solutions

    SUTD, hosting the Asia & Pacific OTIC in Singapore, conducted rigorous security tests with support from Keysight to validate the handover procedures implemented by QCT’s O-RAN-based OmniRAN 5G solution. The testing, involving multiple OmniRAN cells, achieved significant test automation, ensuring consistency and repeatability.

  • Demo 2: O-RAN Energy Saving Demonstration using rApp (from SUTD), OmniRAN (from QCT), and E-Plane Test Suite (from Keysight)

    SUTD showcases its AI-based Energy Saving rApp, designed to manage QCT’s O-RAN-based OmniRAN system optimally. Leveraging AI for traffic predictions and energy saving RAN control, this breakthrough, verified with Keysight’s E-plane test suite, underscores AI’s pivotal role in telecommunications.

  • Demo 3: O-RAN Energy Saving Demonstration using rApp (from SUTD), non-RT RIC (from VMware), and RIC Tester (from VIAVI)

    SUTD presents an integration of its Energy-saving rApp with VMware’s Non-RT RIC and VIAVI’s RIC test solution. Employing state-of-the-art AI technologies, the rApp optimizes cell control, achieving significant energy savings without compromising service quality in realistic scenarios.

MWC 2023

O-RAN security demo

  • Demo 1: Securing Open RAN: Threat-Centric Cyber Validation of QCT OmniRAN

    At Mobile World Congress (MWC) 2023, SUTD, in partnership with Keysight Technologies and QCT, unveiled a “Threat-Centric Cyber-Assessment Framework” coinciding with the launch of the region’s first OTIC in Singapore. The demonstration utilized Keysight’s automation tools to launch severe cyber attacks—such as disrupting time synchronization via DoS and modifying eCPRI packets—against QCT’s OmniRAN infrastructure. Crucially, the tests confirmed that QCT’s equipment remained resilient under these simulated assaults. This successful validation proves the framework is highly effective for securing critical 5G infrastructure, ensuring multi-vendor Open RAN systems are robust, secure, and ready for real-world deployment.

World Expo 2025 (Singapore Pavilion, Osaka)

I coordinated multi-company, cross-country collaboration for an end-to-end 5G Non-Terrestrial Network (NTN) live video-call demonstration over a real GEO satellite link, representing Singapore’s connectivity capabilities at World Expo 2025.

  • Live 5G Satellite Connectivity: Bridging Singapore and Japan via GEO NTN

    This demo showcases a world-first end-to-end 5G New Radio (NR) Non-Terrestrial Network (NTN) connecting Singapore and Japan over a geostationary (GEO) satellite, demonstrated live at the World Expo 2025 Singapore Pavilion in Osaka. The system enables real-time services such as video calls using standard 5G technology, validating seamless interoperability between satellite (NTN) and terrestrial networks (TN).

    Co-developed by SUTD with industry partners, the demo proves that existing GEO satellites can reliably support 3GPP-compliant 5G NR, laying the foundation for global, resilient mobile connectivity and future 6G converged TN–NTN networks for remote and mobility-critical use cases such as maritime and emergency communications.

Research projects

1. Collaboratively train anomaly detection models in edge computing (Research Fellow, SUTD)

  • Research and implementation of a federated learning framework to collaboratively train an anomaly detection (AD) model for edge devices (e.g. multiple robots in different environments belonging to different cooperations).
  • The AD model runs inference at the edge; we implemented federated learning so partners learn a shared model without leaking private training data.

2. Contextual-bandit anomaly detection for IoT data in hierarchical edge computing (Ph.D., SUTD)

  • Adaptive selection among ML models using contextual information from input data for anomaly detection, formulated as a contextual bandit / single-step Markov decision process, optimizing accuracy and delay.
  • Related publications (see also Publications):
    • ICDCS’20 — Mao V. Ngo, Tie Luo, Hakima Chaouchi, and Tony Q.S. Quek, “Contextual-Bandit Anomaly Detection for IoT Data in Distributed Hierarchical Edge Computing,” IEEE ICDCS — Demo Track, Jul. 2020. (Acceptance rate: 30%, 13/43.)
    • AAAI’20 workshop — Mao V. Ngo, Hakima Chaouchi, Tie Luo, Tony Q.S. Quek, “Adaptive Anomaly Detection for IoT Data in Hierarchical Edge Computing,” AAAI Workshop on Artificial Intelligence of Things (AIoT), 2020.

3. Hierarchical mobile edge computing (HEC) for mobile users (Ph.D., SUTD)

  • HEC architecture supporting container migration and base station handover for mobile users. Source code: GitLab repository.
  • Real-world evaluation: Android app and offloaded services (e.g. face recognition with OpenFace, object detection with YOLOv3); sources released publicly.
  • Video demos: YouTube (1) · YouTube (2) · Additional demo (OpenFace migration)
  • GLOBECOM’20 — Mao V. Ngo, Tie Luo, Hieu T. Hoang, and Tony Q.S. Quek, “Coordinated Container Migration and Base Station Handover in Mobile Edge Computing,” IEEE GLOBECOM, Dec. 2020.

4. Soldier Body Area Networks (Research Assistant, SUTD)

  • Dynamic resource management for SBAN.
  • Interference modeling and management for coexistence of single/multiple SBANs and other networks.
  • Energy-efficient MAC for SBAN, with dynamic resource and interference management.
  • Context-aware scheduling for energy efficiency and emergency traffic priority.
  • Heterogeneous wireless integration for roaming and end-to-end QoS.
  • Study of operating band effects on proposed techniques.

Industrial projects

1. IPv6 for APZ system (R&D Engineer, DEK Vietnam)

  • Description: APZ supports multiple IP stack implementations. An IP stack on the Central Processor handles internal I/O communication toward the OSS. This project implemented IPv6 as a layer-3 protocol on the AXE platform with feature parity to IPv4.
  • Skills: TCP/IP, network programming, C/C++.

2. APZ central processor (CP) design (R&D Engineer, DEK Vietnam)

  • Design CP software blocks in APZ for the AXE telephone exchange (BSC/TRC, MSC, HLR, SCP, FNR, TSC, STP, wireline nodes, etc.).
  • Implementation in PLEX-C and ASA-C.