Rhizome: Clustering and Topological Interlocking for the Robotic Assembly of a Martian Habitat
To construct a Mars habitat using complex and irregular Voronoi geometric units, performing unit classification and ensuring proper interlocking assembly is necessary. This paper introduced a robotic assembly method based on the K-means clustering algorithm and Topological Interlocking. The output of clustering and operation analysis can be leveraged for production planning objectives, including resource allocation and scheduling operations for assembling blocks. The topological interlocking aims to develop an inner force-locked system and the processes to design and manufacture the assembly unit. The paper discusses robotic assembly applications of the 3D-printed components of topological interlocking and clustering classification. This research is continued and based on the Rhizome project at the Department of Architecture of the Technical University Delft (TUD) in the Netherlands.
Rhizome Project
This project is a European Space Agency and industry co-funded project focusing on developing an underground Martian habitat. The aim is to develop a materially, structurally, and environmentally optimised 3D printed structure using Robot-Robot and Human-Robot Interaction (R/HRI) supported Design-to-Robotic-Production-Assembly and Operation (D2RPA&O).
Voronoi-Based Components
An autonomous robotic recognition system (ARR), can detect the design components and their properties without placing the robot in a controlled environment. Allowing it to automatically look for every available object in the view or raw materials, and it autonomously estimates the robot posture and target angle for the end effector to grasp.
K-Mean Clustering
The purpose of the K-means in this paper is to cluster the individual Voronoi cells to ensure that the clusters all meet the requirements for assembly. This method prevents the situation where bad clustering could lead to components made of a single or way too many cells.
The overall objective is to create 3D-printable components without support material that are approximately similarly sized and can be assembled into a larger structure. The Voronoi-based logic is inspired by nature and implemented due to its advantages with structural optimization, customizability of the geometry, integration of LSS, etc.
Topological Interlocking
This study introduces topologically interlocking Voronoi-based components, offering customization and efficiency in spatial design at building, component, and material scales. Voronoi-based interlocking patterns provide in the presented study structural stability and strength. The design of the Martian habitat presents a pointed, ogival arch requiring a componential interlocking approach.
Robotic Assembly
Martian Habitat
This method has been tested through simulations, allowing for a clear visualization of the future robot construction process through the stacking of components. With the classification, we can cluster individual Voronoi cells into stackable components. Robots are enabled to handle a large number of different components and their corresponding interlocking. K-means clustering of these cells can segment the parts into clusters based on geometric features or assembly requirements, enabling robots to efficiently plan and execute assembly tasks.