Physics-Aware Multichannel Vector Quantization for Hybrid Digital Twin Modeling of UAV Systems

As Big Data technologies continue to transform industrial manufacturing, digital twin (DT) systems powered by artificial intelligence are becoming key enablers in shaping the emerging low-altitude economy. Among these, unmanned aerial vehicle (UAV) systems play a central role. However, building reliable DTs for UAVs remains challenging due to the heterogeneous nature of sensor data, complex spatiotemporal distributions, and nonlinear interactions within system dynamics. In this work, we propose a fusion framework for UAV-oriented DT modeling that integrates physical priors with data-driven learning, tailored for enhanced DT modeling of UAVs. The method is centered on a multichannel soft vector quantization mechanism, which performs feature clustering across heterogeneous sensor modalities and time scales, enabling robust fusion of multisource data. By embedding physics-aware constraints and leveraging a hybrid fusion strategy, the model enhances generalization under real-world uncertainties. We validate our method by developing a complete UAV DT framework and collecting a large-scale dataset that integrates both simulated and real flight data. Extensive evaluations across hovering, takeoff, tour, and forward flight missions demonstrate superior fidelity and robustness compared to conventional DT models. This work contributes a scalable UAV DT architecture with strong fusion capability and provides open-source tools for advancing self-evolving DT systems in dynamic environments.

• Publication year

  2026

• Venue

  IEEE Transactions on Aerospace and Electronic Systems

• Publication date

  Unknown publication date

• Fields of study

  Physics, Computer Science, Engineering

• Identifiers
  DOI 10.1109/TAES.2025.3639005
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Machine learning-enhanced digital twins for predictive analytics in battery pack assembly

  2025cited by this paper
• Embodied large language models enable robots to complete complex tasks in unpredictable environments

  2025influential reference
• When Embodied AI Meets Industry 5.0: Human-Centered Smart Manufacturing

  2025cited by this paper
• Machine learning for modelling unstructured grid data in computational physics: a review

  2025cited by this paper
• UAVs Meet LLMs: Overviews and Perspectives Toward Agentic Low-Altitude Mobility

  2025cited by this paper
• Digital twin modeling of the robotic gluing system for predicting the quality of glue lines and optimizing gluing parameters

  2025cited by this paper
• A generic digital twin model construction strategy for cross-field implementations with comprehensiveness, operability and scalability

  2025cited by this paper
• An agricultural digital twin for mandarins demonstrates the potential for individualized agriculture

  2024cited by this paper
• A Novel Time Diffusion Model for Industrial Time Series Data Generation

  2024cited by this paper
• Soft Sensor Modeling Based on Vector- Quantized Weighted-Wasserstein VAE for Polyester Polymerization Process

  2024cited by this paper
• Fixed-Wing Unmanned Aerial Vehicle Rotary Engine Anomaly Detection via Online Digital Twin Methods

  2024cited by this paper
• Three-Dimensional UAV-to-UAV Channels: Modeling, Simulation, and Capacity Analysis

  2024cited by this paper
• Data Mode-Related Generative Adversarial Network for Industrial Soft Sensor Application

  2024cited by this paper
• A multi-level digital twin construction method of assembly line based on hybrid worker digital twin models

  2024cited by this paper
• TimeDDPM: Time Series Augmentation Strategy for Industrial Soft Sensing

  2024cited by this paper
• A review of digital twin capabilities, technologies, and applications based on the maturity model

  2024cited by this paper
• A Digital Twin-based Fault diagnostic Method for Subsea Control Systems

  2023cited by this paper
• State-of-Health Estimation With Anomalous Aging Indicator Detection of Lithium-Ion Batteries Using Regression Generative Adversarial Network

  2023cited by this paper
• Digital twin-driven fault diagnosis method for composite faults by combining virtual and real data

  2023cited by this paper
• Straightening Out the Straight-Through Estimator: Overcoming Optimization Challenges in Vector Quantized Networks

  2023cited by this paper
• Development of the Senseiver for efficient field reconstruction from sparse observations

  2023cited by this paper
• CGAN-based synthetic multivariate time-series generation: a solution to data scarcity in solar flare forecasting

  2022cited by this paper
• Direct Data Driven Scheme for UAV Flight Control

  2022cited by this paper
• Digital twin modeling

  2022cited by this paper
• Generative Adversarial Networks

  2021cited by this paper
• Triboelectric nanogenerator sensors for soft robotics aiming at digital twin applications

  2020cited by this paper
• Taming Transformers for High-Resolution Image Synthesis

  2020cited by this paper
• Applying Model-Based Systems Engineering to the Development of a Test and Evaluation Tool for Unmanned Autonomous Systems

  2020cited by this paper
• Conditional GAN for timeseries generation

  2020cited by this paper
• Modelling and simulation of UAV systems

  2020cited by this paper
• Computer science and engineering

  2018cited by this paper
• Wasserstein Generative Adversarial Networks

  2017cited by this paper
• Neural Discrete Representation Learning

  2017cited by this paper
• Learning Structured Output Representation using Deep Conditional Generative Models

  2015cited by this paper
• A Review on Basic Data-Driven Approaches for Industrial Process Monitoring

  2014cited by this paper
• Model-Based and Data-Driven Fault Detection Performance for a Small UAV

  2013cited by this paper

• No citing papers are available for this paper.