⬤ 3D Pose Estimation for Construction Ergonomics: This motion study explores computer vision techniques as a potentially suitable method for data collection in construction. These techniques enable the extraction of 3D pose data from video streams and allow for on-site ergonomic analysis. By implementing these methods, workplaces and physical tasks that may cause excessive physical stress on body parts can be re-designed to create a working environment that accommodates the workers, rather than forcing workers to adapt to the environment.

⬤ Structured Light 3D Scanning: This spatial augmented reality (SAR) research explores a computer-aided visual inspection approach using a projector-camera system for the automated detection and measurements of construction errors during off-site quality inspection. This approach involves using a projector to directly visualize construction as-planned models on the structure currently being built on a site, providing visual guidance to workers and allowing them to compare the as-built objects with the as-planned model.

⬤ Markerless Augmented Reality: This augmented reality (AR) study aims to improve construction and facility management by visualizing an as-planned model on its corresponding surface in the user's view through an AR device (e.g., Hololens) for fast, easy, and correct information retrieval. The markerless AR system does not rely on markers for the localization and registration of scanned environments, thus allowing the user to interactively experience a digital world without the cumbersome tasks of marker installation.

⬤ Immersive Virtual Reality: This study explores the applications of virtual reality (VR) for skills and safety training to understand the immersion-enhancing factors that affect users' experiences and design an interactive interface that can enhance training efficiency. VR simulation provides a realistic and controlled experimental environment where participants' experiences and their psychological responses to environments can be measured in real-time without interrupting the ongoing experience, hardly possible in a field condition.

⬤ Deep Reinforcement Learning: This reinforcement learning (RL) study aims to automate and simulate construction equipment operations for virtual construction by learning operational strategies through trial-and-error experiments in virtual environments. The modeling of an intelligent agent integrates the physical characteristics of actual equipment, enabling the learned agent to emulate its behavior in a field setting. In this context, the RL approach may serve as a foundational research endeavor in the domain of autonomous construction equipment.

⬤ High Dimensional Spatial Data Processing: This study explores a deep learning network for the integration of various geospatial features into the modeling and learning of map-type data (e.g., meteorological and geographical maps). The spatial data processing is inspired by the idea that the data format of digital maps is similar to that of images (e.g., 2D matrix with colors vs. elevations), and thus deep learning techniques can be applied for an understanding and recognition of spatial patterns in a digital map for future predictions.