NIOSH Pilot/Small Project Research Training Grant, 2026-2027, PI: Zixian Zhu
This pilot study aims to (1) identify AI reliance safety risks and (2) assess the effects of AI explanations on worker reliance patterns and decision-making performance. By generating empirical evidence on worker reliance on AI in construction safety, this timely study advances occupational safety research and supports research-to-practice translation by informing prevention-oriented, practitioner-focused guidance and training for the safer deployment of AI-assisted safety tools.
This study proposed a bidirectional communication protocol that integrates natural modalities (gesture, speech, light, sound, and flight paths) to facilitate safe interactions between general construction workers and drones. To evaluate the effectiveness of this communication protocol, both virtual and real construction site scenarios were created for within-subject experiments comparing conditions with the proposed communication modalities against those without them. The study assessed the safety implications concerning physical risks, physiological impacts, cognitive load, and worker perceptions.
Selected Publication:
Zixian Zhu, Ajit Devkota, Idris Jeelani, and Masoud Gheisari. 2026. Bidirectional multimodal worker-drone communication in a real-world construction environment. Elsevier Journal of Automation in Construction, 191: 107171. https://doi.org/10.1016/j.autcon.2026.107171.
Zixian Zhu, Jiun-Yao Cheng, Idris Jeelani, and Masoud Gheisari. 2024. Using gesture and speech communication modalities for safe human-drone interaction in construction. Elsevier Journal of Advanced Engineering Informatics. 62, 102827. https://doi.org/10.1016/j.aei.2024.102827.
This study assessed the physical risks of drones on construction sites using VR-based 4D simulations. A system-level virtual simulation environment was developed to run 2700 simulations per top-risky construction scenario with varying parameters, mimicking and visualizing potential interactions between workers, structures, equipment, and drones. The results suggest that the differential impact of flight parameters and interaction distance on drone collisions implied the need for targeted safety strategies for different working contexts.
Selected Publication: Zixian Zhu, Idris Jeelani, and Masoud Gheisari. 2023. Physical risk assessment of drone integration in construction using 4D simulation. Elsevier Journal of Automation in Construction, 156: 105099. https://doi.org/10.1016/j.autcon.2023.105099.
This study introduces VR-Orient, a mobile application that offers semi-immersive site orientation for smartphones and tablets. The application’s conceptual design was refined through expert panel interviews, resulting in developed content that meets site orientation learning objectives. The technical development of VR-Orient integrates virtual characters, virtual site layouts, and user interfaces that facilitate interaction and reinforce learning on mobile devices. A tablet-based usability study evaluating learning outcomes, system usability, mobile application functionality, informality, and presence indicates promise for advancing construction education, gamified training, and outreach projects, with potential for broader adoption in other labor-intensive sectors.
Selected Publication: Zixian Zhu, Jiun-Yao Cheng, Idris Jeelani, Masoud Gheisari, and Rayment Issa. 2025. VR-Orient: Design and Usability Study of a Mobile Construction Site Orientation Application. Usability and User Experience. AHFE (2025) International Conference. AHFE Open Access, vol 194. AHFE International, USA. http://doi.org/10.54941/ahfe1006663.