As Humans, we have to face a plethora of challenges
in the future, that were primarily caused by our
industrial expansion.
In this course, led by
Prof. Boris Müller, my assigned topic was artificial intelligence.
In search of fictional solutions for todays most
pressing problems with our industrial machinery,
like waste, inefficient manufacturing, unsustainable
energy production, inhumane working conditions and
harmful mining of raw material.
The result mimics the characteristics of plants with
the help of AI to create a self contained, self
evolving and self sustained mechanical organism that
connects buildings for transportation of
commodities, generates energy with sustainable
biogas plants, produces products in fully autonomous
factories and mines raw materials from the ground
like a plant would extract nutrients from the
ground.
To visualise this vision, various layers of root
like tunnels and bud-like factory compartments were
created with a path finding algorithm in Javascript
and imported into the 3d space with a little Python
Script to be animated as a VR Movie to make the
experience of the dimension of this system as
immersive as possible.
An animated rendering to demonstrate the basic principle of the organism.
A Version of the Animation for Virtual Reality Devices
Rendering of single layers of the root structure.
Simplified display of the relationship between the components. On deeper layers, it's mostly about energy and raw material. In the middle, you can see a lot of machine related exchange. The top layer is dominated by consumer related products.
Cross-section of a root. In the middle the transportation of goods happens, while the tubes transport liquid and raw materials before it gets enclosed by a flexible shell.
The expandable factory compartments are ment to transform an input material into an output product. They can store a number of input and output items in their outside compartments, while the processing is taking part in the center.
Factories emerge from small root-branches and can grow and shrink depending on the demand.
To create the roots in a realistic way, I found a pathfinding algorithm to be the most sufficient for the task. Over the course of the project, I wrote a code that generates paths by avoiding defined pixels in a noisemap and export the coordinates of the paths as a csv file to be imported in the 3d application. The green ends of the roots are the base for the factory compartments exported seperately.
The organism is growing comparatively slowly. It's not drilling a hole over a day and is finished like that. It rather grows over decades, constantly adapting to changing demand and conditions. It took this example 50 years from establishment to full size.
Individual consumer access is build directly into the houses or appartments of better situated costumers or shops. For cheaper access, a shared access can be provided to let everyone benefit from the system.
Simple display of the function of the consumer access. It can be used to retrieve goods and service devices from the organsism, or send objects to be stored or processed.
Taken from the D.I.R.T. root database, around 1000 images from various cowpea roots served as inspiration for the project. I also used them to train a machine learning algorithm to generate the roots, yet it didn't lead to the desired result.
A poster to present the result together with the VR Video at university exhibition Werkschau
