Mobile robotic platforms and drones are increasingly being used for supervising construction projects. These technologies, among other things, enable the user to measure clouds of points of the building to represent the progress of construction at any given stage. In the construction industry, one of the primary uses of scanning technologies is to check the 3D digital models of the building developed during the design phase against the actual building itself. The gathered data can thus be used to provide accurate documentation of the as-built structure, the models and drawings of which and their reliability are critical documents throughout the life cycle of the building.

Robots, drones and robotic platforms can be programmed to follow the same route during each as-built survey, so as to gather data and clouds of points in an efficient manner. The route to be followed by the robot can be derived and optimized automatically from the BIM model. Sensors like encoders and IMU units are increasingly present on mobile platforms, and can be used to estimate the relative deviations between different acquisitions of clouds of points.

In our ABC project, Fraunhofer Italia aims to investigate whether this estimation can be used to do away with the need to place visual markers around the site being measured, which are used in manual measurements as points of reference to align the various scans. To this end, regardless of the type of sensors installed on the robot, we have developed an algorithm based on a probabilistic graph which connects the most significant measurements provided by the sensors themselves. The use of a fast mode of iteration between the measured data and the BIM model has the potential to improve advance identification of mismatches in the construction process and thus save time and money.