Autonomous robotic system for on-site waste assessment and separation: Wuhan Demonstration Case

This report presents an autonomous robotic system developed for on-site construction waste assessment and separation within the REINCARNATE project framework. The demonstration focuses on the Longhu Binjiang Tianjie construction site in Wuhan, China, where the system is deployed to support automated identification and handling of construction and demolition waste (CDW). Traditional waste sorting methods in construction environments oftenrely on manual labor, which can be inefficient and potentially hazardous. The proposed robotic system integrates a mobile platform, a lightweight six-degree-of-freedom robotic arm, and a multi-modal perception system combining LiDAR sensing, RGB-D stereo vision, and artificial intelligence–based recognition algorithms. Waste objects are detected using a YOLOv8-based visual detection pipeline and localized through point cloud processing to estimate their three-dimensional positions. The robot can then perform automated grasping and sorting operations using a force-adaptive manipulation strategy. The demonstration results show that the system can effectively support automated waste inspection and handling in semi-structured construction environments. By integrating robotic perception, navigation, and manipulation technologies, the solution contributes to improved construction waste management efficiency, reduced manual labor requirements, and enhanced safety conditions. The system also supports the REINCARNATE objectives of digital waste monitoring and circular construction practices.