Tiny processors, big data, smart computers and alternative realities — the dawning of #Plantech

Tiny Processors

In 1965 Gordon Moore, one of the co-founders of intel, observed how the number of transistors we could fit into an integrated circuit was doubling every year. 50 years later, this trend has continued, and today a typical smart phone boasts processors made up of over 2 billion transistors. This rapid miniaturization of technology accompanied by its increased affordability has resulted in a plethora of experiments of embedding processors and sensors in everything. From lamp posts, concrete structures, cars, parking spaces, rubbish bins, light switches and even toasters and hairbrushes. All of these sensors and processors are also now connecting to each other creating a mesh of nearly 40 billion devices worldwide constantly communicating with each other — what some call the internet of things.

The affordability and size of this technology and the potential to embed sensors throughout the built environment, allows us to build live feedback mechanisms about how people are using cities in a way we have never had before. Monitoring or evidence of Local Plans or Impact Assessments of development proposals, can become significantly more granular and up-to date than it has ever been. Real-time data on the number of homes built, how well they are occupied, the extent to which streets are overcrowded, how often car parking space are being used, or the popularity of parks can all be quantified and measured, and used to inform the policies we write.

By identifying the most impactful data we need to monitor plans and build an evidence base, we can include policies in our plans that require homes to include smart electricity meters, for example, which can provide data to planners about when properties are occupied or vacant and the number of people occupying them.

This real-time understanding of our cities would put pressure on the lengthy and cumbersome plan making process, where policies are often out of date by the time they are adopted. The potential to fine tune, test and iterate policies based on real-time measurable outcomes would become a small step away.

Big data

Passive sensors are only one of the contributors to the 2.5 quintillion bytes of data we produce every day. We actively add to this through the 2.9 million emails we send per second, the 20 hours of videos we upload to YouTube per minute, the 50 million tweets per day, as well as the other plethora of Google searches and Amazon purchases which are continuously tracked.

Despite the numerous new sources of data which we are creating through new technologies, we continue to pay for expensive and time consuming surveys and studies. Consultants

sell rehashed versions of the same data to local authorities and developers over and over again, as evidence base for new plans or impact assessment of proposed developments.

Inadvertently planners have started using new datasets, by consulting satellite imagery and google street view when assessing applications. However, we need to embed datasets from social media, mobile telecoms, internet usage, property transactions amongst the many others into how we monitor and evidence Local Plans, consult the public, and measure the impact of new development.

But collecting data is only half the story. Most of the benefits of data comes from how this data is being used and linked. At Future Cities Catapult we’ve been experimenting with the potential benefits of linking all data from submitted planning applications, viability assessments, SHLAA, and Infrastructure Development Plans, to create a live ‘land portfolio.’ This would provide estimated residual land values for individual sites, social and physical infrastructure capacity estimations and the ability to iteratively test the impact of CIL, Affordable Housing and density policies on site viability.

Smart computers

With the huge growth in processing power and the vast amounts of data we have available to us, we are seeing what many call the dawning of Artificial Intelligence, or as the futurist Kevin Kelly puts it “everything we formerly electrified, we will cognify”. In March last year, Google’s general purpose neural network, DeepMind, learnt how to play the ancient Chinese game of GO. Despite its simple rules, GO has more variables than atoms in the universe, making technically impossible to calculate all potential outcomes. This piece of software not only mastered the game, it beat the South Korean professional Go player Lee Sedol, one of the best players in the world. Similar technologies are already being used to enable driverless cars, image and speech recognition, mark essays, assist with legal advice and even identify some forms of cancer more successfully than clinical pathologists.

Deep neural networks in computers gives us the opportunity to free planners from the many time-consuming laborious tasks planners still have to do today to spend more time on pro-active planning and complex decision making. From the simple automation of basic tasks, such as validating applications, to more complex screening of developments prior to case officer assessments, the potential for this new technology is huge. We are already supporting experiments in automated screening of household planning applications, where a local authority which returns half of 800 applications per year estimates they could save at least 800 hours of processing time if they can reduce this by half through an automated screening process.

We’ve seen Milton Keynes experiment with satellite imagery recognition for planning enforcement, and we’re currently looking into the potential for using image recognition and machine learning to identify development sites for SHLAA’s and Brownfield Land Registers.

Whilst transport modelling is already well used, agent based modelling technology, developed through gaming engines, is giving us the ability to create significantly more complex models. Multiple layers behaviours from population demographics, land market, transport and social infrastructure and even cultural trends can all be modelled simultaneously. Predictive analysis of traffic has moved from transport departments to hand-held devices, and we’re seeing experiments in models that predict areas which are likely to suffer from high crime rates and even gentrification.

Alternative realities

Whilst virtual, mixed and augmented reality aren’t so much of a new technology, it’s only recently that they have become accessible and part of mainstream culture. The use of augmented reality by Pokemon Go or Snapchat filters, the affordability of Google Cardboard as a virtual reality platform, and the mixed reality experience that Microsoft Hololens enables provides us new ways of visualising and experiencing the built environment.

The increasing accessibility and affordability of digital modelling of the built environment through satellite imagery, photogrammetry, LIDAR and the use of drones, means we can easily build virtual reality models of our cities. Models which can be used to explore the cumulative impacts of development on our skyline, to visualise daylight and sunlight and experience the design of our streets and spaces — allowing us to easily iterate and optimise the built form before anything is built.

We are also working with others to explore how the use of augmented reality can be used to experience and visualise proposed developments in-situ, so as well as the much maligned planning notice, people can see simulations of proposals as if they were already built.

What now?

The technological trends above are having a dramatic impact in the way we live, so it is just a matter of time before they fully infiltrate how we plan. Whilst much of it may feel distant, change is inevitable. However, this change needs to happen incrementally, planning authorities should be wary of entering into lengthy contracts with large technology companies. They should focus instead on small and incremental improvements, such as collecting and owning their own data, ensuring that reports and policies are machine readable, and experimenting themselves with new tools and technologies at a small scale.