Lieyun Ding
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Lieyun Ding has a total of 82 co-authors. Here are the top co-authors by number of shared publications:
- Dongrui Wu
Huazhong University of Science and Technology
6 shared publications Cheng Zhou
5 shared publications- Hanbin Luo
School of Civil and Hydraulic Engineering, Huazhong University of Science and Technology, 1037 Luoyu Road, Wuhan 430074, China
5 shared publications Yuyue Gao
5 shared publicationsHanbin Luo
4 shared publicationsYan Zhou
4 shared publications
Latest Publications
- Open Access
AI Agent-Driven Autonomous Operation of Robotic Tower Cranes
Tower cranes are indispensable to building construction, but routine lifting still depends on the coordinated judgement of crane operators and signalpersons, which constrains automation and introduces human-factor variability. This study develops and formalizes an AI agent-driven framework for Level-4 (L4) robotic tower cranes with human supervision. The framework development comprises formal definitions of the digital signalperson agent and digital crane operator agent, a two-phase task–commit protocol, a supervisory finite-state machine, and a maximum-risk decision structure for task acceptance, degraded operation, minimum-risk stopping, and human takeover. Experimental validation was intentionally limited to three operational interfaces: terminal localization, autonomous positioning, and takeover-command latency. Under the tested site conditions, the terminal-localization error had a mean of 3.94 cm and a 90th percentile of 7.11 cm; the autonomous-positioning error averaged 48.2 cm, with 90% of observations within 80 cm; and the takeover-command latency remained below 500 ms, with a maximum of 458 ms. These measurements support the engineering feasibility of the tested supervised-autonomy interfaces, but they do not constitute full validation of the complete L4 safety case, including dynamic risk estimation, obstacle avoidance, adverse-weather operation, blind lifting, or multi-crane coordination.
Open AccessAI Agent-Driven Autonomous Operation of Robotic Tower Cranes
Tower cranes are indispensable to building construction, but routine lifting still depends on the coordinated judgement of crane operators and signalpersons, which constrains automation and introduces human-factor variability. This study develops and formalizes an AI agent-driven framework for Level-4 (L4) robotic tower cranes with human supervision. The framework development comprises formal definitions of the digital signalperson agent and digital crane operator agent, a two-phase task–commit protocol, a supervisory finite-state machine, and a maximum-risk decision structure for task acceptance, degraded operation, minimum-risk stopping, and human takeover. Experimental validation was intentionally limited to three operational interfaces: terminal localization, autonomous positioning, and takeover-command latency. Under the tested site conditions, the terminal-localization error had a mean of 3.94 cm and a 90th percentile of 7.11 cm; the autonomous-positioning error averaged 48.2 cm, with 90% of observations within 80 cm; and the takeover-command latency remained below 500 ms, with a maximum of 458 ms. These measurements support the engineering feasibility of the tested supervised-autonomy interfaces, but they do not constitute full validation of the complete L4 safety case, including dynamic risk estimation, obstacle avoidance, adverse-weather operation, blind lifting, or multi-crane coordination.
- Open Access
- 9 July 2026
- Published by Wiley in Advanced Science
- Vol. 13 (55) , e76488
StackingNet: Collective Inference Across Independent AI Foundation Models
Artificial intelligence (AI) built on large foundation models has transformed language understanding, computer vision, and reasoning, yet these systems remain isolated and cannot readily share their capabilities. Coordinating the complementary strengths of independently developed, black‐box foundation models is essential for trustworthy intelligent systems, yet no established method exists. Here we show that such coordination can be achieved through a meta‐ensemble framework termed StackingNet, which aggregates the output predictions of independent models at inference. StackingNet improves accuracy, reduces individual‐model error and group‐wise disparities, ranks model reliability, and identifies or prunes models that degrade performance, all without access to internal parameters or training data. Across language comprehension, visual attribute estimation, and academic paper rating, it consistently outperforms individual models and classic ensembles, with gains that persist when the base models are uniformly strong. These gains stem from variance reduction and consensus alignment among independent models rather than from any emergent group cognition, and they widen as the model pool grows more diverse. By turning model diversity from a source of inconsistency into a resource for cooperation, StackingNet offers a practical path toward coordinated AI, where progress emerges not only from larger single models but from principled cooperation among many specialized ones.
- 9 July 2026
- Published by Wiley in Advanced Science
- Vol. 13 (55) , e76488
Open AccessStackingNet: Collective Inference Across Independent AI Foundation Models
Artificial intelligence (AI) built on large foundation models has transformed language understanding, computer vision, and reasoning, yet these systems remain isolated and cannot readily share their capabilities. Coordinating the complementary strengths of independently developed, black‐box foundation models is essential for trustworthy intelligent systems, yet no established method exists. Here we show that such coordination can be achieved through a meta‐ensemble framework termed StackingNet, which aggregates the output predictions of independent models at inference. StackingNet improves accuracy, reduces individual‐model error and group‐wise disparities, ranks model reliability, and identifies or prunes models that degrade performance, all without access to internal parameters or training data. Across language comprehension, visual attribute estimation, and academic paper rating, it consistently outperforms individual models and classic ensembles, with gains that persist when the base models are uniformly strong. These gains stem from variance reduction and consensus alignment among independent models rather than from any emergent group cognition, and they widen as the model pool grows more diverse. By turning model diversity from a source of inconsistency into a resource for cooperation, StackingNet offers a practical path toward coordinated AI, where progress emerges not only from larger single models but from principled cooperation among many specialized ones.
- Open Access
- 1 May 2026
- Published by KEAI Publishing in Fundamental Research
- Vol. 6 (3) , 1691-1703
Performance of lunar shell structure for moonbase subjected to low gravity coupled with changing temperature
Temperature variations on the lunar surface can cause significant thermal stress and increase energy consumption within a lunar base. Thus, studying the thermal behavior of lunar structures over time and space is crucial. This study utilizes classical thermomechanical coupling simulations to develop a comprehensive methodology for assessing the thermal performance of lunar shell structures. The model incorporates realistic assumptions about critical loads and boundary conditions and uses the real properties of regolith-based construction materials as input parameters. Using this method, the study simulates 36 scenarios for shell structures, including three types of structures, four latitudes, and three time periods. The temperature field, heat loss, and thermal stress for each scenario are calculated, and the overall trends and specific cases are analyzed. These results can inform further recommendations for the architectural design of lunar shell structures and the selection of construction materials.
- 1 May 2026
- Published by KEAI Publishing in Fundamental Research
- Vol. 6 (3) , 1691-1703
Open AccessPerformance of lunar shell structure for moonbase subjected to low gravity coupled with changing temperature
Temperature variations on the lunar surface can cause significant thermal stress and increase energy consumption within a lunar base. Thus, studying the thermal behavior of lunar structures over time and space is crucial. This study utilizes classical thermomechanical coupling simulations to develop a comprehensive methodology for assessing the thermal performance of lunar shell structures. The model incorporates realistic assumptions about critical loads and boundary conditions and uses the real properties of regolith-based construction materials as input parameters. Using this method, the study simulates 36 scenarios for shell structures, including three types of structures, four latitudes, and three time periods. The temperature field, heat loss, and thermal stress for each scenario are calculated, and the overall trends and specific cases are analyzed. These results can inform further recommendations for the architectural design of lunar shell structures and the selection of construction materials.
- Open Access
- 1 April 2026
- Published by Elsevier in Results in Engineering
- Vol. 30
3D surface design of bobsleigh and luge tracks via performance-based optimization: Lessons learned from Yanqing national sliding center
- 1 April 2026
- Published by Elsevier in Results in Engineering
- Vol. 30
Open Access3D surface design of bobsleigh and luge tracks via performance-based optimization: Lessons learned from Yanqing national sliding center
- Open Access
StackingNet: Collective Inference Across Independent AI Foundation Models
Artificial intelligence built on large foundation models has transformed language understanding, vision and reasoning, yet these systems remain isolated and cannot readily share their capabilities. Integrating the complementary strengths of such independent foundation models is essential for building trustworthy intelligent systems. Despite rapid progress in individual model design, there is no established approach for coordinating such black-box heterogeneous models. Here we show that coordination can be achieved through a meta-ensemble framework termed StackingNet, which draws on principles of collective intelligence to combine model predictions during inference. StackingNet improves accuracy, reduces bias, enables reliability ranking, and identifies or prunes models that degrade performance, all operating without access to internal parameters or training data. Across tasks involving language comprehension, visual estimation, and academic paper rating, StackingNet consistently improves accuracy, robustness, and fairness, compared with individual models and classic ensembles. By turning diversity from a source of inconsistency into collaboration, StackingNet establishes a practical foundation for coordinated artificial intelligence, suggesting that progress may emerge from not only larger single models but also principled cooperation among many specialized ones.
Open AccessStackingNet: Collective Inference Across Independent AI Foundation Models
Artificial intelligence built on large foundation models has transformed language understanding, vision and reasoning, yet these systems remain isolated and cannot readily share their capabilities. Integrating the complementary strengths of such independent foundation models is essential for building trustworthy intelligent systems. Despite rapid progress in individual model design, there is no established approach for coordinating such black-box heterogeneous models. Here we show that coordination can be achieved through a meta-ensemble framework termed StackingNet, which draws on principles of collective intelligence to combine model predictions during inference. StackingNet improves accuracy, reduces bias, enables reliability ranking, and identifies or prunes models that degrade performance, all operating without access to internal parameters or training data. Across tasks involving language comprehension, visual estimation, and academic paper rating, StackingNet consistently improves accuracy, robustness, and fairness, compared with individual models and classic ensembles. By turning diversity from a source of inconsistency into collaboration, StackingNet establishes a practical foundation for coordinated artificial intelligence, suggesting that progress may emerge from not only larger single models but also principled cooperation among many specialized ones.
- 20 November 2025
- Published by Elsevier in Journal of Building Engineering
- Vol. 116
Performance-driven multi-objective optimization for nursing unit design
- 20 November 2025
- Published by Elsevier in Journal of Building Engineering
- Vol. 116
Performance-driven multi-objective optimization for nursing unit design
- 15 July 2025
- Published by Springer Nature in Building Simulation
- p. 1-29
Construct of digital theater for cross-industry BIM data fusion: A scenario-driven ontology-based framework
- 15 July 2025
- Published by Springer Nature in Building Simulation
- p. 1-29
Construct of digital theater for cross-industry BIM data fusion: A scenario-driven ontology-based framework
- Open Access
- 1 July 2025
- Published by Science China Press., Co. Ltd. in Chinese Science Bulletin
- Vol. 70 (21) , 3460-3469
The development trend, objectives, and policy path of lunar base construction
为加快布局月球基地建造领域预先研究,推动我国深空探测可持续发展,本文系统梳理国内外月球基地建造与深空探测的主要发展趋势,揭示我国在月球基地建造中面临的重大机遇与挑战。在此基础上,提出我国月球基地建造的战略目标体系与政策路径建议,为我国月球基地建造的科学规划和决策提供参考。
- 1 July 2025
- Published by Science China Press., Co. Ltd. in Chinese Science Bulletin
- Vol. 70 (21) , 3460-3469
Open AccessThe development trend, objectives, and policy path of lunar base construction
为加快布局月球基地建造领域预先研究,推动我国深空探测可持续发展,本文系统梳理国内外月球基地建造与深空探测的主要发展趋势,揭示我国在月球基地建造中面临的重大机遇与挑战。在此基础上,提出我国月球基地建造的战略目标体系与政策路径建议,为我国月球基地建造的科学规划和决策提供参考。
- 17 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Engineering Management
- Vol. 72 (00189391) , 1163-1177
Collaborative Governance: Blockchain-Based Federated Learning for Construction Safety Service
Deep learning (DL) models are increasingly used to identify unsafe activities for the construction safety service (CSS). However, two typical issues threaten DL training process performance: poor security and privacy in data sharing. To address these problems, a blockchain-based federated learning (BCFL) framework is proposed from a collaborative governance perspective to obtain optimistic DL models for CSS. Two special works of this BCFL framework are that (1) it develops the federated learning empowered privacy-preserving data sharing with the principle of sharing data model weights instead of raw data. Especially for one DL application for CSS, the Fed-YOLOv4 model for workers' unsafe behavior identification task is developed, and (2) it explores two blockchain-based secure model sharing strategies that involve blockchain-InterPlanetary File System combination and model training contribution computing. Then the smart contracts are further developed with the strategies above to streamline the workflow of federated DL model training. Finally, we apply the proposed framework to the practical subway project. The results demonstrate that the proposed framework can improve the DL model training and acquire global DL models for CSS with good accuracy. Our findings indicate that great data security and privacy ensured by introducing blockchain and federated learning can optimize data sharing and then DL models can be improved. Moreover, this study provides managers with a collaborative governance perspective on how DL models can be improved and further applied to CSS. This enables managers to quickly understand project safety performance and make timely managerial decisions for construction management.
- 17 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Engineering Management
- Vol. 72 (00189391) , 1163-1177
Collaborative Governance: Blockchain-Based Federated Learning for Construction Safety Service
Deep learning (DL) models are increasingly used to identify unsafe activities for the construction safety service (CSS). However, two typical issues threaten DL training process performance: poor security and privacy in data sharing. To address these problems, a blockchain-based federated learning (BCFL) framework is proposed from a collaborative governance perspective to obtain optimistic DL models for CSS. Two special works of this BCFL framework are that (1) it develops the federated learning empowered privacy-preserving data sharing with the principle of sharing data model weights instead of raw data. Especially for one DL application for CSS, the Fed-YOLOv4 model for workers' unsafe behavior identification task is developed, and (2) it explores two blockchain-based secure model sharing strategies that involve blockchain-InterPlanetary File System combination and model training contribution computing. Then the smart contracts are further developed with the strategies above to streamline the workflow of federated DL model training. Finally, we apply the proposed framework to the practical subway project. The results demonstrate that the proposed framework can improve the DL model training and acquire global DL models for CSS with good accuracy. Our findings indicate that great data security and privacy ensured by introducing blockchain and federated learning can optimize data sharing and then DL models can be improved. Moreover, this study provides managers with a collaborative governance perspective on how DL models can be improved and further applied to CSS. This enables managers to quickly understand project safety performance and make timely managerial decisions for construction management.
- Open Access
- 1 January 2025
- Published by American Association for the Advancement of Science (AAAS) in Research
- Vol. 8, 0849
Zhuque Base for Martian Habitation: Conceptual Design and Performance Analysis of Cave Dwellings and In Situ Construction
Sustainable habitation systems are critical to enabling deep-space exploration on Mars, where challenges such as low gravity, extreme thermal fluctuations, and resource constraints demand advanced structural innovations. This study introduces Zhuque Base, a habitat concept inspired by terrestrial cave-dwelling principles, optimized through finite element analysis via COMSOL Multiphysics. Three arch-based configurations—the eggshell, catenary, and 2-centered arch—were systematically evaluated by parameterizing geometric variables. The study demonstrated that eggshell arches significantly outperformed 2-centered and catenary arches in mechanical properties, reducing vertical displacement of critical points such as sidewalls by 53% and 44%, respectively. In terms of thermal efficiency, the optimal catenary arch showed 5.5% and 6.7% lower heat loss than eggshell and 2-centered arches. Furthermore, the implementation of optimal parameters (span: 3.2 m, height: 1.25 or 1.45 m) limits the peak compressive stress to 195.72 to 203.38 kPa, while the cross-sectional area can be increased by 14% to maximize the available internal space. These findings establish a parameter-driven framework for in situ Mars habitat optimization, emphasizing the trade-off between mechanical robustness and thermal efficiency in extraterrestrial structural design.
- 1 January 2025
- Published by American Association for the Advancement of Science (AAAS) in Research
- Vol. 8, 0849
Open AccessZhuque Base for Martian Habitation: Conceptual Design and Performance Analysis of Cave Dwellings and In Situ Construction
Sustainable habitation systems are critical to enabling deep-space exploration on Mars, where challenges such as low gravity, extreme thermal fluctuations, and resource constraints demand advanced structural innovations. This study introduces Zhuque Base, a habitat concept inspired by terrestrial cave-dwelling principles, optimized through finite element analysis via COMSOL Multiphysics. Three arch-based configurations—the eggshell, catenary, and 2-centered arch—were systematically evaluated by parameterizing geometric variables. The study demonstrated that eggshell arches significantly outperformed 2-centered and catenary arches in mechanical properties, reducing vertical displacement of critical points such as sidewalls by 53% and 44%, respectively. In terms of thermal efficiency, the optimal catenary arch showed 5.5% and 6.7% lower heat loss than eggshell and 2-centered arches. Furthermore, the implementation of optimal parameters (span: 3.2 m, height: 1.25 or 1.45 m) limits the peak compressive stress to 195.72 to 203.38 kPa, while the cross-sectional area can be increased by 14% to maximize the available internal space. These findings establish a parameter-driven framework for in situ Mars habitat optimization, emphasizing the trade-off between mechanical robustness and thermal efficiency in extraterrestrial structural design.