Metabolism Across Scales | Kuka Endeffector

Metabolism Across Scales focuses on implementing and designing a addition to the kuka endeffector toolbox.

Which can sense and manipulate material units.

The project was conducted as part of the Computational Design

and Digital Fabrication seminar in the ITECH masters program.

Fabian Eidner,Oliver Moldow and Aaron Wagner

The most important part for this project is having a well dialed in 3d Printer that can produce dimensionally accurate parts reliably, apart from that you will need the following:

• Mechanical Endstop (link)
• Digital Servo Motor (Arduino Kit)
• Arduino Uno (Arduino Kit)
• Cheap Webcam (Amazon)
• Soldering Iron
• 6 Jumper wires

Much like the metabolists of the mid 19th century, We regard human society as an evolving process of shaping and reshaping--from atom to nebula. Where we differ though, is in our time’s necessity for reuse and radical environmental change. We do not accept the flawless, stucco arcades that plaster and distract from the potholes in our streets. We believe that buildings are expressions of tiles in their facades. Some may fall and crack, some will spall, materials will change with age and some will be replaced. 

“With Metabolism, the specifically Japanese re-emerges on modern terms. The network becomes the most central urban issue – an organizational form, but also a symbolic image presenting the city and the region simultaneously as structure, organization, and flow, a coherent whole in constant change.”

In the case of the Nagakin Capsule Tower, after 50 years it has been decommissioned and set for demolition. For those on the Japanese board of housing this is a necessary change in the urban ecosystem. However, in our case, its demolition also affords a new methodology of metabolism. Perhaps ironically, the demolition of such a building holds true the mission of the metabolists–that the only constant in construction is destruction and hence, change. 

In our mind, 21st century metabolism does not come from the creation of new cells and therefore structures, but instead reflects the process of revitalizing nodes from previously-used networks in the re-formation of new internet of things… from cells to nebulae. Building components should mold something we call a “circular economy of scale,” in which the built environment benefits from an abundance of prefabricated components… decreasing the embodied energy in every new part manufactured while keeping old parts out of landfills.

In a similar circularity, Cedric Price´s 1980 proposal for the Gilman Corporation was a series of relocatable structures on a permanent grid of foundation pads. Again aligning with the vision of the metabolists, this framework aimed to create a loop between building and user… between building and environment… and finally between building and itself. Here, we are interested in the computation behind the generation of these temporary spaces and also in the automation of reassembly after a space is used.

Because so many of the Metabolists and metabolic-adjacent thinkers have obsessively thought about unitization,we question how metabolism traverses scale and whether or not there is a distinction between picking and placing a cell, a microchip, a block, a chassis, or a housing capsule. We find that what differentiates these might be the democratization of control. Operating at a human scale, like in the case of Nagakin, facilitates a direct translation from disorder to order and creates space at a rapid pace… And these spaces ideally could be created without external intervention from complex, specialized systems and instead be facilitated through automation, where the lay-person is king. 

Otherwise, at larger or smaller scales, it becomes harder to create adaptive, relocatable space… and likewise, create rapid solutions to lack of space--like the vision of Kurokawa. However, there isn’t such a large discrepancy in the logic of placement control. A robot arm is no different from a crane, except in the ability to conformally use geometry as a means for manipulation. Later we discuss utilizing the unique geometry of the capsules as a way to manipulate them.

As Nagakin Capsule Tower is disassembled, we propose to integrate the capsules into our circular economy of building components and constantly re-assemble them into semi-temporary arrangements of space… each iteration keeping in mind the phantom of the one before. In our world, automation will transform the once gridded, cored capsules into freeform networks of housing, resting, and contemplating nodes that are connected with circulation edges that only form from vestigial affordances in the absence of poché.

In order to access the project, we worked at scale and augmentation. The capsules were replaced with boxes, the windows with holes, and the crane with a robotic arm. 

To begin, we developed a gripper tailored to the geometry of the capsules. Instead of gripping from the sides or underneath–which would require a large, complex, and an expensive set of parts, we instead opted for a democratized, compact, and clever implementation inspired by the capsule hatch doors of the ISS that grips from the inside of the window. As the gripper passes the threshold of the window, the movement stops via end-stop and four rotary latches deploy, locking the box in its vertical and horizontal positions. In order to control rotational movement, a small key locks with its respective notch in the window frame.

Simultaneously, we developed our logic of manipulation. Through either user-input or rule-based algorithmic placement, we arrive at a new constellation. The crane or robot then begins its sensing path. It finds the predetermined origin and works backwards, locating all of the boxes (capsules) in space. Once a capsule is detected via edges or QR-code, its plane and orientation are stored in 3D virtual space, later accessed to generate point-to-point movement in KUKAVarProxy. Depending on the next capsule to place, the crane moves above it, offset from the QR-code in its corner and slowly engages with the window frame. After the box has been locked to the gripper, it is then free to access its point-to-point movement from the current location to its site location–either atop a prior capsule or adjacent to an existing one. All the while, keeping in mind the limitations regarding existing entrances, circulation, and windows.

Material Unit:

The appearance and functionality of the material unit was inspired by the reuse of nagakin tower modul geometries but also by the ideas of Cedric Price, that every building component having a computer chip embedded allowing the identification and relocation of the unit.

The nagakin tower modules, a cube structure with a circular cutout as a window. In our case the Unit is Cube 20 x 20 x 20 cm with a 65 mm hole cut out in the middle. Inspired by Cedric Price a unique QR code was used to not only identifying each unit but also allowing for tracking of position and rotation in space.

Hardware:

Having noticed in our last project, that hardware, pun not intended, can actually be pretty hard to debug and make reliably, we opted for a more deliberately simplistic approach.

We went through multiple iterations to get the printed parts to a level at which they were printing very well and also worked as intended (Fig.1).

(Fig. 2) The endeffector itself consists of gripper and actuation gears mounted on a plate which connects them to the servo motor. The top plate is mounted to the connection shaft, where the end stop is also attached. Finally everthing is connected to the mountingplate which also houses the arduino and camera module.

Addons:

Falcon:

Falcon is a Addon for Grasshopper in Rhino 3d which enables the user to sense QR code with unique IDs utilizing a Webcam. The Accuracy is remarkably good, in our case by inputting the right size of the "Marker" we could achieve a accuracy of on average 3mm.

It can be downloaded using the following Link:

https://www.food4rhino.com/en/app/falcon

Firefly:

Firefly is the second mayor Addon used in this project. It enabled us to send commands to the Arduino directly from grasshopper. For using this function a special code has to be uploaded to the Arduino.

It can be downloaded using the following Links:

https://www.food4rhino.com/en/app/firefly

Code / Workflow:

Sequence

The sequencing of our workflow (Fig. 1) starts out with the Robot in its Home position, from here we can go into the sensing pass, effectively scanning the area in a back and forth motion. Each time a Box with its QR is detected the position and rotation is being saved using a Python component. After all the Boxes have been detected we can move on into the pick and place phase. Because of their unique IDs we can start with any box, a user specific, a random or just with the first box followed by an ascending order. The robot moves over the first box and slowly down until the end stop is triggered and thus the gripper being enabled. The Box is then lifted and placed at which point the gripper can be disengaged. If there are still any boxes left we repeat this pick and place sequence until all the boxes are placed. The robot then precedes to again stop at the home position.

Code implementation:

Looking into the flow diagram of the code (Fig. 2) we have three main inputs. The feed of the camera, enabling us to find and track the position of each box using a unique QR code ID, the end stop giving us physical feedback about the height of the box, and finally a safety mechanism using a user input as release toggle