Planning blue-green cities with Artificial Intelligence and Virtual Reality
The sheer complexity of our contemporary cities makes them inscrutable to individual human minds. Urban planning today needs to mobilize collective agency and intelligence to face the challenges ahead.
The DeepGreen project is the result of a design innovation collaboration with the United Nations Development Program. ecoLogicStudio and UNDP have been testing the potential to deploy Artificial Intelligence as part of a new urban green planning strategy.
This strategy analyses hi-resolution data on urban landscape and infrastructure to produce simulated scenarios of sustainable urban development based on recognised biological models. The project has now been formalised as a scalable toolset that can be applied to virtually any city around the world. The simulated urban plans can be visualized by means of powerful hi-resolution drawings, rendered images as well as experienced in immersive Virtual Reality data rooms.
These simulated urban scenarios have three key qualities:
- they are based on extensive hi-resolution urban datasets of the existing city.
- they deploy machine learning techniques known as GAN to simulate urban future re-greening scenarios.
- they generate a visual and morphological output, thus enabling us all to see and interact with the simulated form of our future cities.
In order to appreciate these key features to their full extent ecoLogicStudio had developed a proof of concept Virtual Reality space, as an effective method for interacting with the DeepGreen data ecosystem. In order to allow stakeholders to access the content the VR interface is built on a powerful on-line platform called Mozilla Hubs, which renders widely accessible.
Design research challenge
Currently, more than half of the world's population lives in cities and the urban population is expected to double by 2050. Intense urbanization has led us to rethink human socio-economic development on a global scale with particular emphasis on the relationship between human beings, their footprint and the environment. Cities and city regions today are at the forefront in both feeling the effects and fighting to offset the potentially catastrophic effects of climate change.
Cities are the biggest CO2 emitters globally, and therefore it is necessary to redesign their infrastructure, rethink consumption patterns and circularity. In other words, it is necessary to find ways to convert what cities expel, as waste or pollution, into raw material to feed new processes of production.
This entails innovative strategies of waste management, water conservation and recycling, renewable energy production and trading. It also involves implementing technologies for the filtration and re-metabolisation of air pollution. At the same time, we recognize that in each urban space there are layers of informality which supplement and complement existing public services, whether it is in water catchment, individual waste recycling, or in other forms such as decentralized construction. These dynamics, while an essential part of the tapestry of life in cities, are not always recognized. Yet, effective ways of addressing vulnerabilities demand utilization of the entire spectrum of human and environmental systems in cities.
We can design resilient cities that use their size and collective energy to create refuge for both humans and displaced wildlife, that promote the emergence of positive microclimate, that replenish depleted water sources and that restore degraded terrains, pushing back on processes such as desertification, land erosion and contamination. This entails innovative strategies of urban re-greening and re-wilding as well as of urban agriculture. A critical quality of urban planning today it to mobilize collective agency and intelligence to face the challenges ahead. In this way local solutions can be evolved in response to the given challenge.
To achieve this goals Deep Green has been designed as a planning solution combining the scalability of a sophisticated planning application to the design sensibility and intuitive accessibility of its design interface. This enables a high degree of customisation and evolution to each specific urban application, or urban design solution.
At its core, Deep Green uses sophisticated algorithms to analyse hi-resolution data on urban landscape and infrastructure (mostly available open source) to produce simulated scenarios of sustainable urban development. Crucially it enables a new approach to green urban design, one that is driven by a deeper understanding of biological systems and models, is iterative and adaptive to multiple scenarios globally.
As demonstrated by our test run with early adopter cities such as Aarhus, Tallinn, Barcelona, Caracas, and with UNDP partner cities Vranje, Guatemala and Mogadishu, our design scenarios simulate the evolution of restorative urban networks.
Re-wilding Guatemala City
In Guatemala City this scenario is exacerbated by a serious lack of waste management. The Guatemala city garbage dump is the biggest landfill in Central America containing over a third of the total garbage in the country. 99% of Guatemala's 2,240 garbage sites have no environmental systems and are classified as "illegal." Deep Green enables us to design beyond traditional planning concepts such as zone, boundary, scale, typology and program. It recognises the true nature of contemporary cities as complex dynamical systems.
Guatemala City is situated on a complex and highly unstable terrain surrounded by mountains and volcanoes, some of which are still active. Its ecosystems, originally very rich in biodiversity, are now made fragile by unchecked urbanisation and, given its climatic zone, the effects of climate change. Our approach creates an interface between bottom-up processes of self-organisation, such as the many local waste recycling activities that are emerging out of necessity in the areas closer to the dumping sites, and the strategic decision making that occurs at municipal, national and international level.
The aim is to find new synergies and direct investments where and when have they have the most potential to engender positive change. Two proposals emerged in this analysis of Guatemala City: a re-wilding plan foster a new coexistence between human and urban wild animals, and an urban agriculture plan proposing a method to guarantee food security and to employ the impoverishes rural population currently migrating to the city of Guatemala. Both proposals are sensitive to local conditions while effecting international power relationships.
For instance, the migrating birds populating the green areas of Guatemala City migrate to and from Canada. Therefore, investments in urban re-wilding will have benefit in the biodiversity of Canada. Migrant workers cross the city on their way to the US-Mexico border. Urban agricultural plans could retain rural workers alleviating the pressure on both Mexico and the US. Such synergies have the potential to channel significant international funds to local projects improving the life of citizens of Guatemala City.
Similarly in Mogadishu, another case study, land degradation is a key environmental issue and is closely related to desertification, drought and unsustainable livestock and agricultural practices. The problem is exacerbated by a completely horizontal development of the city and the critical lack of water. Vegetation is so sparse that its restorative effect becomes negligible.
In our proposal a bottom-up water collection and filtration network is computed to optimize its performance in relationship to existing urban density and road networks connectivity. This is then reinforced by a large scale re-greening strategy aim at creating dense networks of plants in proximity to the areas of collection, thus promoting the emergence of restorative microclimates and ecological niches. Crucially here the algorithmic interface allows the simultaneous computation and optimisation of contrasting parameters and the development of a multi-scalar approach.
The urban re-greening strategy has therefore multiple hierarchies. Locally it recognises gaps in the urban vegetation and gives guidance for the planting of trees in optimal locations. At the scale of neighbourhood is optimises the location of water collections points serving the existing buildings. At the urban scale it fosters densifications of the vegetated network to promote the emergence of ecological niches and local microclimates, especially around water collection zones. At the territorial scale it promotes the emergence of a barrier, natural and man-made, to push back desertification and restore some of the abandoned agricultural plots as well as the infrastructural networks of canals and water wells.
Vranje Renewable Energy City Region
The third scenario of this study is the city of Vranje in Serbia. It is a very different kind of urban system compared to the two previously described.
Vranje is a distributed city region, with a small size and little resident population but with far greater opportunities in terms of establishing a new regional network economy. The study recognises opportunities latent in the surrounding territory in the production of renewable energies from sources such as solar, wind, hydraulic and biomass. Such locally distributed production could give rise to an emergent and integrated renewable energy network capable of generating significant new circular economies and high-level employment opportunities.
In conclusion, with Deep Green we are developing an urban design methodology and related technology that effectively deploys Artificial Intelligence to the re-greening of global cities and city regions. The method is scalable and we intend to develop a dedicated urban design application to drastically increase the number of cities that could benefit from this technology.
In order to perfect it we seek to apply it to the design of several more test cities with clear urban design challenges. This design research and innovation project has been developed with the support of the Innovation Facility funding, and is a part of the City Experiment Fund, supported by the Slovak Ministry of Finance.