About

Hubo Cai is a Professor and Associate Head of Civil and Construction Engineering at Purdue University, located at the Lyles School of Civil and Construction Engineering. His research interests include construction engineering and management, built environment modeling with BIM and GIS, engineering database and information management systems, virtual construction, and sensor and sensing technologies in construction automation. He is actively involved in advancing the understanding and development of construction processes and technologies, contributing to the fields of construction engineering and management through his research and academic leadership.

Research topics

• Computer Science
• Artificial Intelligence
• Computer Security
• Machine Learning
• Construction engineering
• Engineering
• Human–computer interaction
• Psychology
• Geography
• Civil engineering
• Risk analysis (engineering)
• Operations management

Selected publications

• Pocket LiDAR for Dimensional Measurement of Precast Concrete Panels

  Open MIND · 2026-04-14

  otherOpen access

  This study evaluates the feasibility of pocket LiDAR as a low-cost solution for dimensional quality control (QC) of precast concrete elements. Traditional methods rely on manual tape measurements, which are labor-intensive and prone to human error. High-precision technologies such as terrestrial laser scanning (TLS) are expensive and difficult to operate. To bridge this gap, this research assesses pocket LiDAR based on comparisons with TLS and tape measurements through controlled experiments. Point cloud data of a concrete slab specimen were collected in both pre-casting and post-casting stages using TLS and an iPhone LiDAR system. The datasets were evaluated in terms of point cloud fidelity, dimensional accuracy, and compliance with industry tolerances. Results show that pocket LiDAR achieved a root mean square error (RMSE) of 5.5 mm in rebar position measurements compared to TLS, significantly outperforming tape measurements (45.8 mm RMSE). For concrete dimensional measurements, pocket LiDAR demonstrated comparable accuracy to tape measurements. Compliance analysis further indicated that pocket LiDAR can reliably identify most tolerance violations in rebar position and concrete dimension, although minor discrepancies were observed in borderline cases. Overall, pocket LiDAR demonstrates strong potential as a practical alternative for rapid and automated QC in precast construction.

  PublisherDOI

• Pocket LiDAR for Dimensional Measurement of Precast Concrete Panels

  Zenodo (CERN European Organization for Nuclear Research) · 2026-04-14

  otherOpen access

  This study evaluates the feasibility of pocket LiDAR as a low-cost solution for dimensional quality control (QC) of precast concrete elements. Traditional methods rely on manual tape measurements, which are labor-intensive and prone to human error. High-precision technologies such as terrestrial laser scanning (TLS) are expensive and difficult to operate. To bridge this gap, this research assesses pocket LiDAR based on comparisons with TLS and tape measurements through controlled experiments. Point cloud data of a concrete slab specimen were collected in both pre-casting and post-casting stages using TLS and an iPhone LiDAR system. The datasets were evaluated in terms of point cloud fidelity, dimensional accuracy, and compliance with industry tolerances. Results show that pocket LiDAR achieved a root mean square error (RMSE) of 5.5 mm in rebar position measurements compared to TLS, significantly outperforming tape measurements (45.8 mm RMSE). For concrete dimensional measurements, pocket LiDAR demonstrated comparable accuracy to tape measurements. Compliance analysis further indicated that pocket LiDAR can reliably identify most tolerance violations in rebar position and concrete dimension, although minor discrepancies were observed in borderline cases. Overall, pocket LiDAR demonstrates strong potential as a practical alternative for rapid and automated QC in precast construction.

  PublisherDOI

• Knowledge-Based Digital Inspection System for Training Construction Inspectors

  Transportation Research Record Journal of the Transportation Research Board · 2026-05-07

  articleSenior author

  State transportation agencies face growing challenges in training construction inspectors owing to a shrinking experienced workforce and the increasing complexity in inspection tasks. Traditional document-based training methods are often fragmented and lack contextual depth, limiting their effectiveness in preparing novice inspectors. This study presents a knowledge-based digital inspection training system that consolidates inspection information from multiple Indiana Department of Transportation (INDOT) documents—including Standard Specifications, General Instructions to Field Employees, Indiana Testing Methods, standard drawings, and Manual for Frequency of Sampling and Testing and Basis for Use of Materials—into a semantically structured knowledge graph. The inspection training system enhances learning through the integration of rationale, instructions, construction pitfalls, and failure scenarios associated with inspection tasks. A user-centered web application was developed to deliver this content in an intuitive format aligned with how inspectors naturally perform their duties—by pay item, construction process, or risk scenario. The system was evaluated through a mock exercise involving INDOT inspectors. Results showed that the system is effective in improving the construction inspectors’ confidence and understanding of inspection rationale, instructions, and risk consequences of missed inspection. This research contributes a scalable, risk-informed framework that improves accessibility and comprehension of inspection knowledge, with the potential to foster proactive inspection behaviors and support more consistent construction quality control across transportation projects.

  PublisherDOI

• Uncertainty-aware localization and orientation estimation of underground pipelines from GPR data

  Advanced Engineering Informatics · 2025-11-22

  articleSenior authorCorresponding
  PublisherDOI

• Automated and Semantic-Rich BIM Modeling of MEP Systems Using Monocular Cameras

  Journal of Computing in Civil Engineering · 2025-07-16

  articleSenior author

  Accurate and updated as-built building information modeling (BIM) has become an effective tool for managing mechanical, electrical, and plumbing (MEP) systems due to its capabilities in enhancing visualization and data interoperability. Video-based 3D reconstruction methods offer a cost-effective and portable solution for creating as-built BIM models. However, the process of extracting images from video often results in low-quality input data due to issues such as blurriness and distortion, and the low accuracy and density of reconstructed point clouds results in challenges for segmentation and detailed modeling. Using monocular cameras only, this paper presents a novel, automated BIM modeling approach that overcomes these challenges. The newly designed method consists of three compartments: an adaptive frame extraction method that efficiently generates high-quality image sequences from video; the key image identification and semantic transfer techniques that simplify semantic segmentation by converting it into a 2D task on selected images; and the undirected graph-based modeling approach combined with prior knowledge that can produce semantically rich BIM models. The method was tested across multiple scenarios. Results demonstrated its reliability in generating accurate as-built BIM for MEP systems from videos. The method fills an important gap in automated BIM modeling to provide accurate as-built information for managing MEP systems.

  PublisherDOI

• User Manual for Digital Inspection Training System

  2025-01-01

  report

  Construction inspection plays a pivotal role in ensuring the quality and long-term performance of infrastructure products. INDOT is currently facing the challenge to effectively train novice inspectors and update experienced inspectors with the latest versions of quality requirements to perform construction inspection effectively and more efficiently. This project developed a digital inspection training system to train INDOT inspectors to inspect asphalt pavement construction. Inspectors can access the inspection instructions through a construction process-based approach or a pay item-based approach. The inspection guidance is developed using data and information from INDOT’s Standard Specification (2024), Manual for Frequency of Sampling and Testing and Basis for Use of Materials, General Instructions to Field Employees, Standard Drawings, risk-based check items developed in previous JTRP projects (SPR 4002 and SPR 4422), INDOT training materials, and relevant external references from Federal Highway Administration, state transportation agencies, and reports of National Cooperative Highway Research Program (NCHRP) projects. The system is implemented through a Web platform. It serves as a training tool for novice inspectors and a quick reference for experienced inspectors seeking update-to-date specifications and guidelines.

  PublisherDOI

• An Ontology-Based Framework for Controlling Robotic Excavator Using Goal-Oriented Natural Language

  2025-12-11

  articleSenior authorCorresponding

  Technical advancements in robotic excavators are leading a trend towards human-robot cooperative earthmoving work. A well-known example is the teleoperation of excavators. Traditionally, the robotic excavator performs the actions by reading human intents directly from joystick control signals. However, this process heavily relies on the sensorimotor capability of skilled operators in completing complex joystick operations, which may result in high human workload and is error-prone. Communicating through goal-oriented natural language is a more natural way to fully utilize robot intelligence to achieve harmonious cooperation. This paper proposes an ontology-based framework of goal-oriented natural language for the control of robotic excavators. The framework is composed of four modules: task understanding, knowledge management, task and motion planning, and sensing and perception. A simulation prototype in Unity3D is developed to conduct a preliminary validation of the proposed framework. The findings are expected to enable the control of robotic excavators through goal-oriented natural language instructions.

  PublisherDOI

• Advances in lower hybrid technology at 4.6 GHz towards long-pulse tokamak operation

  Plasma Science and Technology · 2025-10-20

  article

  Abstract To extend the long-pulse operation at higher plasma current and plasma density on Experimental Advanced Superconducting Tokamak (EAST), a new lower hybrid current drive (LHCD) system operating at 4.6 GHz with 4 MW nominal power is being under development. This paper reports the recent advances in lower hybrid technology including the key components of the klystron, launcher, and circulator, which are supported by the Comprehensive Research Facility for Fusion Technology (CRAFT) project. A prototype klystron at 4.6 GHz with one output window has produced 500 kW radio frequency (RF) power with a duration of 1590 s on the test bed with matched load. An active cooling launcher based on the passive-active-multijunction (PAM) concept at 4.6 GHz has been constructed, which is waiting for the plasma experiments on EAST. For the prototype circulator, a full power (namely, forward power = 500 kW) operation over 1135 s with 30% power reflected back to itself has been achieved. These achievements make a large step towards the possible implementation of an LHCD system with the same frequency of 4.6 GHz on China Fusion Engineering Test Reactor (CFETR).

  PublisherDOI

• Underground Utility Network Completion based on Spatial Contextual Information of Ground Facilities and Utility Anchor Points using Graph Neural Networks

  Proceedings of the ... ISARC · 2024-05-27 · 1 citations

  articleSenior author

  Underground Utility Network Completion based on Spatial Contextual Information of Ground Facilities and Utility Anchor Points using Graph Neural Networks Yuxi Zhang, Hubo Cai Pages 936-943 (2024 Proceedings of the 41st ISARC, Lille, France, ISBN 978-0-6458322-1-1, ISSN 2413-5844) Abstract: Every year, accidental damage during excavation leads to numerous disruptions in utility services. These incidents cause not only financial losses but also injuries and fatalities. A major contributing factor is the lack of accurate location data for these utilities of such incidents. The current practice involves a time-consuming coordination process of obtaining utility maps from owners and field surveys, which is often hindered by delays and incomplete records. In response to these challenges, this paper proposes a novel method to predict underground utility lines in situations where records are unavailable or delayed. Our approach leverages visible utility anchor points, such as manholes, and the spatial context provided by nearby ground features like roads. The methodology involves three primary steps: constructing a relational data model of the utility network, transforming this data into graphs, and employing a graph neural network for prediction. This innovative approach demonstrates good performance, achieving a 94.12% ROC AUC score in predicting sewer line connections between manholes. This method automates the inference of utility lines, providing utility owners and excavation contractors a solution for identifying unknown connections and reducing risks from inaccurate information. Keywords: Underground Utility Network Completion, Spatial Contextual Information, Graph Neural Networks DOI: https://doi.org/10.22260/ISARC2024/0121 Download fulltext Download BibTex Download Endnote (RIS) TeX Import to Mendeley

  PublisherDOI

• Human Language-Instructed Robotic Excavation based on Behavior Trees

  AHFE international · 2024-01-01

  articleSenior author

  Collaborative construction robots have emerged as a promising alternative to relieve construction workers from both physically and cognitively demanding tasks, contributing to a safer and more productive construction industry. However, communicating with robots is not a trivial task as human workers and robots speak different languages. From the human-centered perspective, allowing human workers to communicate with robots using natural language is desirable because it minimizes additional cognitive load to human workers. Existing studies, however, have been focusing on converting language instructions into sequential actions, leading to a rigid task plan and inability to handle complex situations and unstructured working environments. To address this critical limitation, this paper explores the use of behavior tree (BT), an alternative architecture for describing and controlling complex tasks like excavation. A behavior tree is a hierarchical tree structure that specifies the switching between the agent’s actions (i.e., execution nodes) via control flow nodes. Its modular nature allows the BT of excavation to be generated through linking reusable actions based on the human task descriptions. The resulting BT structure enables the robot to alter its behavior by selecting different tree branches in response to changing working conditions, thus improving its adaptability to dynamic construction environment and its capability of error-handling. In addition, the BT eases the human understanding of robot behavior for debugging and correcting robot behavior. A corresponding framework is proposed for enabling humans to guide a robotic excavator using goal-oriented language instructions. The framework consists of four modules: interpretation and reasoning, knowledge management, structural analysis and parsing, and BT generation. The interpretation and reasoning module decomposes instructions into structured executable intents. The knowledge management module organizes the knowledge for instruction reasoning, including the robot capable skills and its current working environment. Structure analysis and parsing module further grounds the intents and extracts associated parameters, while BT generation module maps the extracted elements with predefined BT nodes, building and refining the BTs of desired tasks. A case illustration is performed to demonstrate the viability of the proposed framework with executable demos. The findings are expected to facilitate efficient and transparent human-robot cooperation in earthmoving construction from a human friendly perspective.

  PublisherDOI

Recent grants

• Mapping the Underworld of Buried Utilities - A Hybrid Sensing Approach

  NSF · $217k · 2015–2019

• Collaborative Research: Correlating Geospatial Data Lineage and Positional Accuracy for Excavation Damage Prevention

  NSF · $150k · 2013–2016

Frequent coauthors

• William Rasdorf

  North Carolina State University

  56 shared

• Osama Abudayyeh

  41 shared

• Jiannan Cai

  The University of Texas at San Antonio

  40 shared

• JungHo Jeon

  Seoul National University

  39 shared

• Steven J. Fenves

  39 shared

• Robert O’Neill

  Florida Gulf Coast University

  38 shared

• Kincho H. Law

  Stanford University

  38 shared

• Pingbo Tang

  Carnegie Mellon University

  36 shared

Education

• PhD, Department of Civil, Construction, and Environmental Engineering

  North Carolina State University

  2004

Awards & honors

• Civil Engineering Alumni Achievement Award
• Construction Engineering Outstanding and Emerging Leader Alu…
• LSCCE Distinguished Engineering Alumni