Platform-independent visual installation progress monitoring for construction automation

Efficient interior progress monitoring is crucial for the timely completion of construction projects. Although robots have been used for data acquisition to automate interior progress monitoring, existing methods do not adequately consider the variations of robot platforms and different types of construction environments, which makes it challenging to apply these methods to various robots and environments. This paper proposes an integrated system that achieves automated interior installation progress monitoring, which can be applied to various construction environments and robot platforms. Algorithms are proposed to systematically generate navigation goal points for robot navigations based on BIM information of objects, which enables the detection of target objects with the proper viewing distance and angle. A transformer-based object detector is used to recognize the installation status of building elements, and a progress updating module is developed to correlate the detection results with the robot’s sensory and BIM information to generate a construction progress report for interior installation. This framework hierarchically estimates the percentage of project completion and allows for tracking of the installation work progress. The proposed system has been verified through laboratory and onsite experiments using various platforms, including a mobile robot, a four-legged robot, a drone, and a smartphone camera.