Abstract
Logistics in construction sites represents one of the main causes of musculoskeletal disorders for workers in the long run. For this reason, logistics represents a task that is highly suitable for automation. However, the application of robotics in construction is hindered by the on-site environment, which is often unstructured and subject to frequent change. To tackle this problem, the system proposed in this paper is a collaborative robotic platform aimed at human-centered applications on-site. The platform follows an operator while it carries heavy loads, such as materials and equipment, and it stops when it is near to the assisted worker. The system does not only avoid obstacles that are detected by its sensors, but it also navigates thanks to its knowledge of geometric and semantic information of the building project. This is achieved through a bridge between the Robot Operating System and the project data contained in the Building Information Model. The proposed system was developed and tested in laboratory, where it followed an operator while avoiding dynamic obstacles and areas marked in the building project file. This system represents a step forward towards the application of on-site collaborative construction robots.