Please, note that this report has been submitted to the EC but its content has not been approved yet. Deliverable D1.2 of the DISCOVER project (HORIZON-CL4-2023-TWIN-TRANSITION 01-11-101129909) presents the development and implementation of Navi-Wall, an open-source navigation software for autonomous mobile manipulators tasked with non-invasive data acquisition in buildings designated for demolition. Navi-Wall is a core component of Work Package 1 (WP1), which focuses on creating the Oliwall robot—a mobile manipulator equipped to scan and identify construction materials and hidden elements in walls, floors, ceilings, and columns using advanced sensors such as LiDAR, GPR, hyperspectral cameras, and RGB-D vision. The deliverable describes the methodology and results associated with robot navigation and data collection strategies. It includes algorithms for 3D environment reconstruction, semantic mapping, global and local path planning, and wall-docking manoeuvres for high-precision scanning. The software architecture integrates ROS 2 control frameworks, custom planning strategies, and reinforcement learning methodologies to support future integration of surface and subsurface navigation. The approach is validated in both simulated and real test environments using an equivalent mobile robot platform. The report documents the facilities, sensor setup, software components, and control strategies developed, highlighting the modular design and adaptability of the Navi-Wall system. The work lays the foundation for autonomous, accurate, and safe material identification workflows in circular construction and deconstruction processes.
Universitat Politècnica de Catalunya (Thu,) studied this question.