Mickey Matter’s research focuses on the development of a highly cost-effective production method, using a compression moulding process. Custom aluminium moulds in three different scales were designed and fabricated using CNC milling. These can then be used to produce cheap, but highly precise pieces. The material used is ABS pellets, but other materials such as plaster-composites and flexible plastics were also tested. The elements are designed in two parts, with a joint connection. This allows to make the elements hollow and lightweight.
Using a custom-made vacuum gripper attached to the industrial robotic arm, the building blocks can be picked up by the robot and assembled with a simple, rapid pick and place mechanism. The spherical geometry reduces the need for high-tolerance and precision in the assembly. The rounded surface help the elements slip into place, which increased the speed. The elements can be picked up by a vacuum-suction gripper, and then put into place.
The team developed a computational method based on combinatorics that is able to efficiently assemble these building blocks into complex, functional structures. The algorithm tests different combinations of the elements, defines connection points and helps to explore different design possibilities. The building blocks are combined with each other in different rotations, producing a variety of patterns. These can then be evaluated in terms of connection strength and robotic fabrication constraints. Afterwards, the data is sent to a robot for assembly.
To prototype the assembly process, the team of students fabricated two chairs and a table. These small-scale prototypes help to develop precise rules for larger-scale, architectural assemblies. The first chair and the table are generated with a symmetrical bias. A limited amount of combinatorial patterns is used, resulting in a highly controlled design. The second chair explores an assembly with less constraints, resulting in an asymmetric, less controlled form.
On an architectural scale, Mickey Matter believes that their system can be used to robotically assemble large building elements in a factory environment, which would then be transported and assembled on site. The architectural elements could be fabricated with a range of materials such as concrete, timber, plastics etc.
The Bartlett School of Architecture’s BPro Research Cluster 4 (RC4), lead by Gilles Retsin, Manuel Jimenez Garcia and Vicente Soler, develops design methods for robotic fabrication. In previous years, RC4 has experimented with 3D printing, using industrial robots. These processes were tested on furniture pieces, such as a plastic chair (http://www.dezeen.com/2016/02/05/bartlett-students-ucl-3d-printed-filigree-chairs-robots/ ) and a concrete table (http://www.dezeen.com/2016/01/21/amalgamma-develops-3d-printing-concrete-technique-building-structures-bartlett/ ). However, these 3D printing processes remain time-consuming, and have difficulties with multi-materiality. Moreover, if a mistake occurs during the printing process, the whole object has to be printed again - there is no “undo-button”. 3D printed objects are not reversible, and can not be recycled easily.
A new generation of research initiated by RC4 moves into Discrete Robotic Assembly, rather than 3D printing. Essentially, this process assembles lego-like building blocks into complex forms. Just like Lego’s, these standardised building blocks are always the same. Rather than using the robot to craft a complicated form with hundreds of different elements, the complexity here emerges from the combination of simple building blocks. These building blocks can be understood as “voxels” or volumetric pixels - a digital method which is popular in computer visualisations for scientific analysis. Discrete Robotic Assembly is inherently faster than 3D printing and other forms of robotic assembly. It reduces the cost, allows for different materials and at the same time maintains a high level of formal complexity. These properties open the possibility for robotic fabrication to scale up to an architectural scale. A first iteration of research has been tested again on the scale of furniture. Projects include building blocks made of timber, compression moulded plastics, metal wires and extruded plastic.
MickeyMatter : robotically assembled furniture
B-Pro Design Computation Lab - Research Cluster 4, the Bartlett School of Architecture, UCL
Tutors: Gilles Retsin, Manuel Jiménez García with Vicente Soler