Adaptive Robotic Construction

This project, developed for the Neura Robotics Challenge, explores how autonomous mobile robots can adaptively locate, pick, and place construction-scale construction elements in response to live spatial conditions. The system integrates a holonomic mobile platform, a seven-axis collaborative robot arm, an industrial structured-light 3D scanner, and a real-time computational design environment into a closed-loop robotic assembly workflow. Rather than relying on fixed work cells or pre-taught trajectories, the robot scans its environment, localises available bricks, sends this information into Grasshopper, and receives back a task-specific sequence of motions generated from the parametric design model.

By linking mobile autonomy, 3D perception, and live design-driven motion generation, the project demonstrates an alternative to conventional teach-and-repeat robotics in architecture and construction. The system can absorb small variations in platform position, material location, and site conditions through software rather than mechanical recalibration. In doing so, the design model becomes not only a representation of the intended assembly, but an active source of executable robotic behaviour, pointing toward more adaptive and site-responsive approaches to robotic construction.

CREDITS

  • Johannes Braumann,
  • Karl Sigline