| description abstract | Regular bridge inspections are key to maintaining healthy infrastructure and to preventing unanticipated structural failures. In recent years, mobile inspection robots have been proposed as a tool to aid bridge inspections. Their key advantages include the ability to access areas of a bridge that are otherwise difficult to inspect and the ability to automate the process of collecting and processing data to increase repeatability and reliability of inspections. Although software algorithms have successfully demonstrated the ability to automate defect detection, many data collection platforms, such as drones and ground vehicles, still require an operator to control the robot via teleoperation, and they do not adequately address the challenges associated with automating data collection for bridge inspection, which is central to achieve repeatability of inspections over time. In this study, the challenges associated with automating data collection for visual inspection of bridges are addressed using a ground-based robot, and an autonomy framework, which can meet the requirements for management and execution of inspection plans, is presented. Key tasks considered in this study are managing inspection plan execution on a ground-based robot, robot localization and mapping, and autonomous navigation in a bridge environment. This automated data collection framework is demonstrated on a concrete bridge by automatically building an accurate point cloud reconstruction of the bridge. Automating data collection can enable more systematic and repeatable inspections, which is a critical upstream task to constructing deterioration models in structural components using inspection data collected over the lifetime of a bridge. | |
