Human-Centered Automation for Resilience in Acquiring Construction Field Information

Resilient acquisition of timely, detailed job site information plays a pivotal role in maintaining the productivity and safety of construction projects that have busy schedules, dynamic workspaces, and unexpected events. In the field, construction information acquisition often involves three types of activities including sensor-based inspection, manual inspection, and communication. Human interventions play critical roles in these three types of field information acquisition activities. A resilient information acquisition system is needed for safer and more productive construction. The use of various automation technologies could help improve human performance by proactively providing the needed knowledge of using equipment, improve the situation awareness in multi-person collaborations, and reduce the mental workload of operators and inspectors.

Unfortunately, limited studies consider human factors in automation techniques for construction field information acquisition. Fully utilization of the automation techniques requires a systematical synthesis of the interactions between human, tasks, and construction workspace to reduce the complexity of information acquisition tasks so that human can finish these tasks with reliability. Overall, such a synthesis of human factors in field data collection and analysis is paving the path towards “Human-Centered Automation” (HCA) in construction management. HCA could form a computational framework that supports resilient field data collection considering human factors and unexpected events on dynamic job sites.

This dissertation presented an HCA framework for resilient construction field information acquisition and results of examining three HCA approaches that support three use cases of construction field data collection and analysis. The first HCA approach is an automated data collection planning method that can assist 3D laser scan planning of construction inspectors to achieve comprehensive and efficient data collection. The second HCA approach is a Bayesian model-based approach that automatically aggregates the common sense of people from the internet to identify job site risks from a large number of job site pictures. The third HCA approach is an automatic communication protocol optimization approach that maximizes the team situation awareness of construction workers and leads to the early detection of workflow delays and critical path changes. Data collection and simulation experiments extensively validate these three HCA approaches.