Prefabricated timber construction offers major productivity benefits, yet on-site assembly, while largely machinic, remains human-controlled and ad-hoc. This paper presents an integrated multi-scalar construction framework for automated on-site assembly of point-supported cross-laminated timber floor slabs. Precise placement of large timber panels with cranes is currently hindered by occlusions, limited sensing, and unreliable pose verification during final alignment. The proposed system combines a digitally controlled tower crane with compact distributed robots embedded into timber elements. Each robot performs onboard vision-based detection of inherent building system features and transmits observations to a central coordinator, which fuses multi-view data to estimate the panel's pose. The fused estimate guides crane motion during placement, while passive alignment features enable force-guided final insertion. Large-scale outdoor experiments demonstrate reliable markerless multi-camera pose estimation and mechanical self-alignment. The results indicate that distributed robotic perception can enable robust crane-based assembly under realistic site conditions. Toward autonomous timber construction using distributed robotic perception system in coordination with a tower crane. • Multi-scalar construction framework coupling tower crane and distributed robots. • Integrated methodology toward autonomous on-site timber construction assembly. • Distributed robotic perception system performs local, online vision-based sensing. • Centralized fusion of local perceptions estimates continuous panel pose. • Field experiments demonstrate robust distributed vision under site conditions.
Opgenorth et al. (Fri,) studied this question.