Gardens of Things — The use of IoT to monitor and maintain biodiverse urban green spaces
Aude Vuilliomenet is a PhD Student at The Bartlett, Connected Environment Lab, Center for Advanced Spatial Analysis (CASA), UCL. Her research “Gardens of Things” focus on urban green infrastructure, outdoor IoT, deep learning on tiny computing devices. She is currently working with volunteer communities to build and install electronic devices to monitor urban gardens in real-time. Her approach aims to offer novel insights of the ecological health of the gardens as well as to explore how digital tools facilitate community engagement with nature and maintenance of biodiverse green spaces.
Introduction
Natural habitats are havens for wildlife, facilitators of flood mitigation, and playgrounds for human activities. In an era of increased urbanisation, green spaces can help cities to adapt their built environment to tackle a range of challenges.
Mayors around the world are embracing nature-based solutions to make their cities greener and more resilient as well as to achieve a Net Zero Future. Ahead of the COP26 and following the release of the IPCC report, a number of institutions have released guidebooks providing frameworks on how to deliver sustainable cities as well as declarations pledging their commitment to take actions. See for example “The Green City Accord: Speeding-up Europe’s Green Transition” or “C40 Nature Declaration”.
Yet, throughout the world there lacks an appreciation for the integration of remote sensors to maintain and augment green spaces. My experiences have positioned myself at the intersection of the built environment, urban informatics and food systems (read more about my background here), and my aim is to apply my specialist knowledge to undertake groundbreaking research in the application of data-driven technologies within urban green initiatives. In this article, I explored how the rise of vertical farming and the miniaturisation of electronics has shaped the idea behind my “Gardens of Things” research project.
The importance of green spaces
Urban green spaces have gained particular attention over the past decade. This interest peaked sharply during the COVID-19 pandemic where citizens turned towards open green spaces, parks, forests, and gardens. While institutions closed their doors, natural areas became an excellent substitute for keeping social contacts and staying physically active (Kumar V. and Vuilliomenet A.). Moreover, the marked increase and alarming severity of recent extreme weather events drew attention to the potential of green infrastructure to tackle the climate crisis. Architects, urban planners, and policymakers are pressed to provide design solutions to mitigate the harms caused by heatwaves, heavy rains and floods, but also to find answers to improve waste management, air quality, ecological biodiversity, and food security. Today’s dialogue on the role and place of nature in cities is however nothing new.
In the introduction of “Dense+Green Cities: Architecture as Urban Ecosystem”, Thomas Schröpfer revisits the development of urban gardens: from the hanging gardens of Babylon to the engineered parks of Paris, stopping by Howard’s vision of “Garden City” as well as James Corner’s New York Highline. The multiplicity of examples highlight the complex relationship between nature and urbanism, and especially point out the harsh competition for space. Thus, Schröpfer advances the use of greenery into the vertical dimension of cities.
Vertical Farms
The idea of green densification or the integration of plants in the three dimensional space may sound familiar to vertical farming enthusiasts. The concept of using the vertical space to grow food was considered eccentric in the early 2000. Twenty years later, cities all over the world have embraced it and called it the solution to “Feeding the world in the 21st century”, referring symbolically to the title of the book written by Dickson Despommier, the credited father of vertical farming.
The principle behind vertical farms is to produce food indoor, without soil, and in urban centres, allowing for a more energy and resource efficient system: food is produced locally reducing food transport miles; the growing environment is fully controlled independent of climate conditions and eliminating pesticide, fungicide or insecticide dependence; multi-stacked shelves optimise the need for space; and, nutrients and water delivery are carefully calculated and automated to compensate the need for fertile soil.
Empty urban spaces such as roof-tops, abandoned parking lots and other high-rise buildings in metropolitan areas have been the most fertile breeding grounds for the evolution of urban farming systems. Multiple ingredients are necessary to their success. Development in artificial lights, rainwater collection, nutrients recycling and growing medium have all a fair bit to play in producing high quality and flavourful crops. But, the secret at the heart of the productivity of these farms has really been and continue to be the advancement in microcontrollers and single-board computers.
AI-powered IoT Devices
The profusion of ultra-low-power, low-cost embedded devices has allowed rapid automation and control over the growing process. It provides the possibility to collect a tremendous amount of environmental data points to feed machine learning algorithms in order to create the perfect conditions for each crop to grow. Similarly, computer vision and cloud computing allows the analysis of thousands of images to identify variations in the size, width, colour, curvature of the growing plants. The health of crops are thus maintained and any potential yield loss is limited. The partnership between AeroFarms and Nokia Bell Labs AI, unveiled in August 2021, illustrates only the beginning of collaborations between vertical farming, agricultural businesses and deep tech companies.
The augmentation of vertical farms with AI-powered IoT devices is fascinating. However, for cities to deliver the resilience, liveability and sustainability that our world needs, we need to consider the outdoor urban green spaces. This is why my “Gardens of Things” research is looking at how these tiny electronics devices can be used to sustain urban vegetation as well as present tangible environmental metrics of the integration of greenery in the built environment.
Smart Urban Nature
Nature is becoming smart. The embedding of IoT devices in natural spaces can now track birds, monitor trees, count bees or even listen to bats. These “Nature Smart” projects have been an inspiration to formulate my PhD. For instance, London’s “Shazam for Bats” was developed to capture the soundscape in Queen Elizabeth Olympic Park. The recorded acoustics data were processed at the edge and a convolutional neural network (CNN) was applied to identify bat calls and the related bat species. “My Nature Camera” is another example of wildlife monitoring. This project was designed so that everyone could build their own nature camera, and thus photograph the wildlife surrounding them. Moreover, Treemania provides an end-to-end solution for the management of trees. From a moisture sensor to real-time data, it allows urban gardeners to make informed decisions on when to water trees. Similarly, ApisProtect has developed its own technology solution to help beekeepers taking care of their hives. All of these examples belong to what Nadina Galle has defined as the “Internet of Nature”; the underlying idea is that the digitalisation of green spaces will open opportunities to understand the role of urban greenery on ecological activities. Landscape architects, urban gardeners and other citizens can now make data-driven decisions to enhance urban ecology.
Smart Urban Nature (1) “My NatureWatch” Camera (2) Interior of EchoBox showing Intel Edison and Ultrasonic Microphone (3) EchoBox in Queen Elizabeth Olympic Park by (naturesmartcities.com)
TinyML for Urban Gardens
There is huge, unexplored potential to integrate AI to learn from the natural environment in order to maintain and monitor urban green spaces. I am only getting started with my PhD research, but I am particularly excited to explore the rapidly growing field of TinyML or, in other words, running deep learning models on tiny, low-power chips. Thanks to the book written by Pete Warden and Daniel Situnayake and the work of the TensorFlow Lite team, it is now possible to integrate embedded devices into spaces of the urban realm that are affordable and energy efficient. As rightly said by Daniel Situnayake “There is some magic in the idea of TinyML”, I am eager to work out the potion that would provide insights into the biodiversity and activity of our urban green environment.
