Scholarlink

<p>Digital Twin (DT) has been increasingly implemented to improve reliability, transparency, and decision-making in industrial robotics. The MObile MAnipulator (MOMA), which combines an autonomous mobile robot with a collaborative manipulator, provides high mobility and manipulability to meet the growing demand for flexibility and agility in intelligent manufacturing. Although the heterogeneous robots in MOMA perform different functions, they are driven by a shared energy source. Therefore, the Prognostics and Health Management (PHM) of MOMA remains challenging, stemming not only from the coupled dynamics of navigation-manipulation functions but also from the heterogeneity of sensing modalities involved. To address these challenges, this paper proposes a DT-based endurance prediction and health monitoring approach for heterogeneous MOMA. First, a systematic architecture, including the physical layer, DT layer, machine learning layer, and dashboard layer, is proposed, and the intra- and inter-layer data flows are defined in detail. Subsequently, a continuous bidirectional telemetry data pipeline is established to consolidate heterogeneous real-time data from the physical MOMA and maintain data synchronization between physical and virtual spaces. A data-driven prediction method based on the CatBoost model is employed to provide remaining useful cycle estimates, enabling mission-level endurance assessment. Furthermore, a virtual reality-assisted health monitoring dashboard is developed to support human-in-the-loop execution supervision and reactive control for MOMA operations. Finally, the proposed approach is experimentally validated with a MOMA in the laboratory. The results demonstrate the proposed framework is a practical solution for MOMA PHM in intelligent manufacturing.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Digital Twin-Based Endurance Prediction and Health Monitoring for Heterogeneous Mobile Manipulator

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月10日

查看全文
  • Yunhan Zhang,
  • Xiaolang Yang,
  • Khurshid Aliev,
  • Dario Antonelli

摘要

Digital Twin (DT) has been increasingly implemented to improve reliability, transparency, and decision-making in industrial robotics. The MObile MAnipulator (MOMA), which combines an autonomous mobile robot with a collaborative manipulator, provides high mobility and manipulability to meet the growing demand for flexibility and agility in intelligent manufacturing. Although the heterogeneous robots in MOMA perform different functions, they are driven by a shared energy source. Therefore, the Prognostics and Health Management (PHM) of MOMA remains challenging, stemming not only from the coupled dynamics of navigation-manipulation functions but also from the heterogeneity of sensing modalities involved. To address these challenges, this paper proposes a DT-based endurance prediction and health monitoring approach for heterogeneous MOMA. First, a systematic architecture, including the physical layer, DT layer, machine learning layer, and dashboard layer, is proposed, and the intra- and inter-layer data flows are defined in detail. Subsequently, a continuous bidirectional telemetry data pipeline is established to consolidate heterogeneous real-time data from the physical MOMA and maintain data synchronization between physical and virtual spaces. A data-driven prediction method based on the CatBoost model is employed to provide remaining useful cycle estimates, enabling mission-level endurance assessment. Furthermore, a virtual reality-assisted health monitoring dashboard is developed to support human-in-the-loop execution supervision and reactive control for MOMA operations. Finally, the proposed approach is experimentally validated with a MOMA in the laboratory. The results demonstrate the proposed framework is a practical solution for MOMA PHM in intelligent manufacturing.