End-Effector Pose Estimation and Control for 3D Printing with Articulated Excavators

We introduce a novel large-scale autonomous mobile manipulator system based on an instrumented and automated construction machine for precise on-site sensing and fabrication. The system is based on an automated hydraulic walking excavator equipped with IMUs and LiDAR units. In this work, we develop the technology to precisely map, localize, and move the printhead in the environment. By fusing GNSS localization with kinematic sensing of the mobile machine and end-effector, we get a globally consistent and locally accurate positioning for in situ robotic construction printing process. Moreover, we present a control approach that enables the excavator to move the end-effector precisely along predefined trajectories. We evaluate the performance of the proposed system in a variety of real-world tests in the field and analyze different sensor modalities and arrangements. Finally, we discuss the potential applications, including the fabrication of non-standard architectural forms and increased safety.