Background
Due to the lightweight character of cables as driving components, cable robots can span very large workspaces or produce highly dynamic movements of the robots’ tool centre. Exemplary applications that utilize these system properties are the Skycam (L.L. Cone, 1985) for television broadcasting or the ETH field phenotyping platform at the Swiss Federal Institute of Technology (N. Kirchgessner et al., 2016) in Zurich, which is suspended over a field and is used for continuous monitoring of the planting. At Max-Planck-Institute in Tübingen, a completely restrained cable robot was realized for high-dynamic motion and flight simulation (P. Miermeister et al., 2016). Beside the mentioned applications, there are various works, that demonstrate the suitability of cable robots for applications in construction industry. Such as, automated mounting of facade elements (Itturalde et al., 2020), or hybrid (additive and subtractive) manufacturing of concrete elements (M. M. Muniz et al., 2020) and recently, automated building of masonry (Boumann et al., 2020). Due to the intrinsic modularity and flexibility of the cables, cable robots offer the unique possibility of a reconfiguration option. Thus, by reconfiguring the system, the properties of the cable robots can be adapted to changed requirements. The authors Nguyen et al. (Nguyen et al., 2014) and Gagliardini et al. (Gagliardini et al., 2015) demonstrate in their works how reconfiguration can be used for large-scale aircraft maintenance and sandblasting of large structures.
Problem statement
Based on the aforementioned works, we identified a gap in research, we want to address in this project, that is given by a comprising reconfiguration approach, that extends from the planning of a geometry according to requirements to finally, a machine learning based calibration method which bridges the gap from the planning stage to an economical use of the reconfigurable cable robot.
Goals
The overarching goal of RP27 is to establish a reconfigurable cable robot as a large-scale and fully automatable robotic system for the use in a dynamic construction environment. The integration of the reconfigurable cable robot in the IntCDC cluster enriches the spectrum of robotic systems, which extends from small-scale robots with a single driven joint to large-scale automated cranes, by a robotic system whose features were not previously available. This allows for new cross-system cooperation approaches that open up entirely new fabrication and construction approaches for Co-Design. Hereby, reconfigurable cable robots can contribute with very large workspaces combined with full controllability of six degrees of freedom. To enable seamless digital and physical integration into the IntCDC infrastructure, the first step is to define the requirements in collaboration with the IntCDC project partners. Based on that, this project will develop a digital designing method that determines an optimal cable robot geometry for the regarded use case. In order to show the practical suitability as well as optimize operational properties of the robot, there will be carried out experiments to generate a data basis to learn the kinematic parameters of the cable robot and feedback them for online calibration of the kinematic model.
Get in touch
Felix Trautwein
M.Sc.Research Assistant "Mechatronic Systems and Processes"