Announcements

We are delighted to invite the winner of the Buildings Best Paper Award, Prof. Dr. Eric Jing Du from the University of Florida, and his co-authors, PhD student Hengxu You and postdoctoral researcher Dr. Tianyu Zhou, to discuss the paper “Robot-Enabled Construction Assembly with Automated Sequence Planning Based on ChatGPT: RoboGPT”. In this interview, Prof. Dr. Du shared his insights on the award-winning paper and cutting-edge research directions, provided guidance for the long-term development of the journal, and offered profound perspectives on academic research.

Award-winning article:
“Robot-Enabled Construction Assembly with Automated Sequence Planning Based on ChatGPT: RoboGPT”
by Hengxu You, Yang Ye, Tianyu Zhou, Qi Zhu and Jing Du
Buildings 2023, 13(7), 1772; https://doi.org/10.3390/buildings13071772

Author profiles:
Dr. Eric Jing Du is a professor in the Department of Civil and Coastal Engineering at the University of Florida. Prior to his academic career, he served as a senior data analyst at Zachry Industrial. Dr. Du earned his Ph.D. in engineering management from Michigan State University (2012) and a bachelor’s degree in civil engineering from Tianjin University, China (2004). His current research focuses on developing next-generation intelligent information systems for future civil engineering projects. Hengxu You is a Ph.D. student on Dr. Du’s team, and Dr. Tianyu Zhou is a postdoctoral research associate on the team.

Photo of authors Tianyu Zhou and Hengxu You with the award certificate

Research background:
The origins of this paper trace back to late 2022, when ChatGPT first emerged. Professor Du initiated a lab discussion on integrating ChatGPT with the team’s ongoing research. The core goal was to explore whether large language models (LLMs) could understand and execute specific physical tasks—using an LLM as the robot's “brain” to decompose complex tasks. This later became the prototype of the VLA model.

1. Could you please briefly introduce the main content of the award-winning paper?
This paper studies LLMs’ ability to understand and perform specific tasks, such as pipe connection or object placement. We first parameterized the simulated environment’s scene and input it into the LLM, which then generated step-by-step operation instructions based on its reasoning. Next, we developed a “translator” to convert these instructions into robot control signals. This approach used the LLM as the robot’s “brain” for task decomposition and planning. At a time when LLMs lacked specialized reasoning capabilities, the results exceeded expectations, achieving a task success rate of about 84% in repeated experiments. Of course, such outcomes are more common now with subsequent advances in reasoning models.

2. What are the future research frontiers in this field? Any upcoming projects?
Frontiers include full automation, human–robot interaction in civil engineering, and AI’s practical support for construction workers. We are at a critical turning point in AI development—shifting from pattern recognition in information and data to true understanding and feedback in the physical world. This trend is referred to in the literature as “Embodied AI” or in industry as “Physical AI.” Traditional AI focuses on pattern recognition, classification, and prediction, while emerging research emphasizes deep understanding of the physical world, task planning, and execution. Thus, directions like agentic AI, physical AI, and embodied AI are deeply integrating AI with physical systems. There are significant opportunities in vehicles, intelligent transportation, aviation, robotics, etc., with potential for large-scale system automation in the future.
However, we also face major challenges, especially regarding a severe lack of data. In civil engineering, data is often collected longitudinally, and cross-industry data-sharing mechanisms are still underdeveloped. Therefore, we believe one key future direction is promoting “data governance”—systematically collecting and integrating domain data based on consensus, which could lead to breakthroughs.

3. Any advice for young researchers?
From my perspective, young researchers often focus on specific metrics and short-term, easily publishable topics—this is understandable, as everyone needs to gradually build their academic credentials. However, as you grow, I suggest calming down to set longer-term goals. Build your research system step by step, like building blocks, ensuring each phase of work continues to add value over the next five years or more. So, for young scholars like PhDs and assistant professors, at a certain stage, you must learn to step back from immediate tasks and think deeply about where the field will be in 10–15 years and what role you want to play. Then work backwards to plan each step. Such planning is crucial.

4. How do you feel about receiving this award?
Hengxu You mentioned that we are both surprised and delighted by this honor and deeply grateful for the academic recognition. This achievement truly belongs to the entire team—my co-authors and senior colleagues all made invaluable contributions. A special thanks to Professor Du, whose early vision of integrating GPT with robotic arms demonstrated remarkable foresight, preceding industry trends by several months. His guidance was instrumental in bringing this paper to fruition.

5. Buildings was a relatively new journal at the time. What attracted you to submit your article?
We realized this research direction would attract broad attention in civil engineering and other engineering applications, and the angle was relatively novel, so we wanted to publish the results promptly. We learned that Buildings promised a fast review cycle and rapid indexing, and the journal focuses on building-related technologies, which aligned well with our research topic. Therefore, we completed and submitted the paper as efficiently as possible.

Finally, a sincere thanks to Professor Du’s team for taking time out of their busy schedules to be interviewed by the Buildings Editorial Office and for openly sharing their research journey and valuable insights. Thank you, Professor Du, for your support and suggestions for the journal’s development, and thanks to the award-winning authors for their explanations and analysis of the professional field. We sincerely wish the team continued success and more groundbreaking results on their future research journey!