AI Photosynthetic Organism in the Metropolitan area of Barcelona

AI Photosynthetic Organism is a study in machine learning as a highly informed methodology for urban and rural farms design.
The project was developed in relation with our Internet of Cities project with the overall aim to increase the carbon Capture in the city of Barcelona. Our strategy focuses on incrementing the efficiency of the land to capture CO2 and is based on a system of three interconnected elements, working on three different scales. For the distinct goal of this course two of them were further studied.

System description. U-Pick farms, Rural Farms, Urban Farms

Reference: Ekket, ReGen Villages

Rural Farms

At first we focus on the rural areas of the periphery of Barcelona. We used Circuitscape software in order to simulate the animal flows, which is a fundamental mechanism to connect the different areas and make them work as a living organism.
Thus, using neural 2D style-transfer we understand how the system is going to evolve depended on the natural flows of animals. We experimented in combining the anthropological structures with natural patterns in relation with the concept of converting the crops to organically organized structures.

Rural Farms. Large scale experimentation, speculative system, CO2 capture

Rural Farms - Step 1. From anthropological to organic structure

Rural Farms - Step 2

Rural crops system evolution. 2D Style transfert, deep learning


U-Pick Farms

A key element to promote the change is the u-pick farms. We locate them in relations with train connections to create a network of sustainable agricultural production, that serves the communities of farmers and consumers. Their role as educational nodes is fundamental for the success of our whole system.
In order to capture more carbon in the crop areas, we developed a genetic algorithm. The algorithm takes into consideration the expected productivity along with the symbiotic behavior per plant and calculates their spatial distribution.

U-Pick Farms. Downscale experimentation, reactivation elment, generative crops

The results of this process were calculated for the area of our study. Later this output was combined with some images of existing permaculture farms in order to have a visual realistic output.
Later, the output was combined with the terrain of our site (DEM file) and using z-brush we managed to obtain a realistic 3D representation of our case of study, located in san Vincent del’s Horts.

U-Pick farms - Step 1. Crop genetic algorithm.

U-Pick Farms step 1. Style transfert 1.

U-Pick Farms step 1. Style transfert 2.

U-Pick Farms. 3D rendering

This experience enabled us to distinguish the varying tools in machine learning. As designers we can understand these technologies as new tools which provide us a new paradigm on how to develop new frameworks for our project aims.


Urban AI | Carbon capture Barcelona is a project of IAAC, Institute for Advanced Architecture of Catalonia developed at Master in City & Technology in 2020/21 by student: Kevin Aragón,  Iñigo Esteban, Diana Roussi, Tugdual Sarazin and faculty: Prof. Sandra Manninger & Dr. Matias del Campo