Bridge inspection traditionally relies on scaffolds, snooper trucks, and aerial robots for hard-to-reach areas beyond the visual line of sight of inspectors. This study automates the collaborative inspection workflow for multiple users when inspecting hard-to-reach bridge sections by interfacing low-cost multimodal sensors mounted on a magnetically wheeled climbing robot with augmented reality devices in a cyber-physical system. Non-destructive and visual sensors provide near real-time information of automated steel thickness measurements and visual observations of corrosive regions, respectively. The visual sensors also support robot teleoperation. The cyber-physical system is evaluated for its user-friendliness and performance. The system performance is quantified by the live streaming lapse and the execution time for non-destructive testing in augmented reality. Audio notifications are effective for simple training instructions as they help complete 77.8% execution tasks in 0.33 minutes, while video demonstrations are preferred for complex instructions to complete 88.9% tasks in 0.29 minutes. The video captured from a GoPro mini camera is received by the inspectors with approximately 1 second delay when wirelessly transmitted to the host, while it takes 0.122 seconds over cabled transmission.
Runji et al. (Thu,) studied this question.