Research summary report of C06

Integration of Additive Manufacturing in the Construction Process

[07.05.2026]

Savadkouhi, Mohammad; Doctoral researcher, m.savadkouhi@tu-braunschweig.de

Mawas, Karam; Doctoral researcher, k.mawas@tu-braunschweig.de

Maboudi, Mehdi; Associated scientist, m.maboudi@tu-braunschweig.de

Gerke, Markus; Project leader, m.gerke@tu-braunschweig.de

all: TU Braunschweig, Institute of Geodesy and Photogrammetry (IGP)

The main goal of C06/IGP

Following a comprehensive study of geometric quality control approaches and stages for integrating additive manufacturing (AM) into construction, the second phase of C06/IGP focuses on integrating AM into a cyber-physical construction system (CPCS). This system enables bidirectional interaction between virtual and physical environments, supporting continuous digital workflows and automated feedback loops. Therefore, the first goal of C06/IGP is to evaluate reality capture sensors in terms of measurement accuracy and level of automation to identify optimal solutions for quality control. The second goal is to develop a near-real-time localization workflow for localizing dynamic on-site members, such as robots, sensors, and humans.

Summary

A CPCS may consist of multiple interacting physical members, including printing robots, quality inspection sensors, and human workers, some of whom may be equipped with augmented reality (AR) devices. On the one hand, selecting the optimal quality inspection sensor is crucial, as its output directly influences how accurately the physical process is captured and how it can be integrated into the digital system for real-time monitoring and control. On the other hand, a fundamental requirement in a CPCS is the precise and up-to-date spatial localization of all interacting members. In addition to information such as fabrication design, construction progress, and the status of system components, reliable localization is critical for ensuring consistent coordination between the physical and digital environments. Together, these challenges highlight key research needs in sensor selection and real-time localization for the effective integration of AM into a CPCS.

Therefore, several reality capture sensors, namely a terrestrial laser scanner (TLS), a professional photogrammetry system, an industrial structured light scanner (SLS), as well as a handheld SLS, are employed to capture multiple medium-sized 3D concrete-printed objects with varying surface geometries. This experiment provides comprehensive information regarding the accuracy and resolution of 3D modeling for each sensor, as well as the applied co-registration strategies, required manual effort, and flexibility of each reality capture technique, among others.

In addition, to localize the aforementioned members of the CPCS in real time, a motion capture (MoCap) system is employed at the Digital Construction Site (DCS) which serves as the main research infrastructure of C06 in the second phase. As an initial experiment, the MoCap system is used to track the end effector of a large-scale gantry-type shotcrete 3D printing robot (SC3DP) at the DCS (see Fig. 1). Furthermore, research is being conducted on the localization of reality capture sensors for direct registration, as well as on the real-time navigation of mobile robotic systems.

Current state of research

The experiment on optimal sensor selection has been completed by capturing 3D models of five specimens using three reality capture techniques: laser scanning, structured light scanning, and photogrammetry. The objects of interest comprise one 3D-printed concrete box with multiple geometric features, two paddle-shaped specimens, and a pair of positive-negative T-shaped joint specimens. The outputs of the sensors are compared with respect to resolution, noise, geometric deviation, etc.

Moreover, with a focus on the accuracy of end-effector positioning, the end-effector of the large-scale gantry-type SC3DP robot at the DCS is tracked under both static and dynamic scenarios using motion-tracking-based localization. In the former, the deviation is computed at predefined fixed locations, whereas in the latter, it is evaluated along pre-planned paths. The final output is a 3D deviation map of the entire printing area, providing a foundation for future compensation strategies for real-time correction of end-effector position.