A Bird’s Eye View of the Smart City
I had the opportunity to visit the “Smart Cities & Communities” event in Manchester recently, where a number of key speakers eloquently presented their view of the smart city – what it might be and what might help shape it.
It is easy to pick out a few lead ideas and rapidly develop tunnel vision in the excitement of exploring certain sexy technologies at the macro level but instead, I’d like to ride that metaphorical helicopter and take a bird’s eye view.
Why? Because I went to the event with the aim of identifying the key technologies, how they work, who’s developing them and possibly understanding how to get involved – but I came away with something unexpected.
This had nothing to do with the technologies involved. My team work with most of them regularly - just a simple matter of understanding which ones are winning. They work in all the normal ways and due to the scale of the endeavour the vendors are, on the whole, big or well-established players such as Nokia, ARM, NEC, Dassault and Blue Wireless – to name but a few.
But what I’d like to share with you here is how and why a few well thought out, well presented and admirably conveyed messages had to stop me in my tracks and make me think of things from very different perspectives.
The first one might be obvious to many people. It’s about who holds our data and what is done with that data. The smart city will be filled with sensors monitoring everything from energy use to water quality. This sort of data is impersonal, monitored on smart lamp posts or within building infrastructures. This is useful to a degree and few would object to its use but even more useful is personal data – location, mode of transport, speed of transit, Wi-Fi usage, health data etc… The point was made that we all carry around powerful sensor platforms capable of passing on this information – this is “citizen produced” data. Both streams of data are likely to be combined into “urban sharing” database platforms with the aim of alleviating congestion, deploying services efficiently, reducing energy and the like. The fact is that to make it really, REALLY useful, some of this data must be made publically available so that citizens can make informed choices as they pass through or interact with the smart environment. But who will hold the data? Who can we trust and why? Who will balance openness of the data with extracting monetary value and who will ensure that our trust (and privacy) is not abused? These questions, of course, are not new either but the scale of it might be - remember that I am flying over the city looking down at all the people, all the cars, all the buildings and all the lamp posts and suddenly, the enormity of the data sets hits me. Hold that thought.
Before I move on to a different aspect of trust, let me convey one final aspect raised by a member of the conference audience. Say for a moment that we are successful. Data sets are aggregated and analysed. Citizen data is “re-distributed” openly. Everyone knows how best to navigate the city and its services. EVERYONE. The best way in, the fastest way out, where the most people are… A great point was made in that use of data with evil intent does not make the data itself evil. That’s true and applies equally to the internet of today. Maybe we shouldn’t allow this to worry us - I’ll leave you to decide. Either way, the General Data Protection Regulation is being re-vamped to help cope with this brave new data world, and in 2018 will address issues such as consent, transparency, profiling, data transfer and breach notification. Cyber security, of course, extends beyond our personal data and into the vulnerability of the infrastructure itself. It is heartening to see this being taken seriously at parliamentary level – we ignore it at our peril.
The last aspect of trust that I thought was interesting was regarding trust in the technology itself. From my bird’s eye view I see autonomous vehicles travelling effortlessly along shiny roads, taking optimum routes and delivering stress free and accident free commuting. And then I’m jolted out of my Utopia. Two fundamental observations from two different speakers do the trick:
First, “Smart cities will not be as painted in illustrations by graphic artists – cities are crowded, messy and difficult places”.
The introduction of the smart city will be slow and faltering and we need to be realistic about what we expect from our technology. Staggeringly, a UN study predicts that 70% of the world’s population will live in cities by 2050 so we cannot expect technology to alleviate overcrowding per se. Rather, technology could help manage it or make it a little more bearable – a little less difficult. Work is being done, particularly in London, to examine how re-organisation can help achieve a far better commuting experience. They are imagining a city infrastructure more akin to a radio mesh network, where travel between A and Z can take multiple routes through equally well serviced hubs – essentially de-centralisation. With the ability to change direction at any hub, freedom of movement could mimic a data packet in a fault-tolerant system. If successful, free movement could be guaranteed to such an extent that one major reason for opting for personal transport is removed.
But there is another challenge here. Let’s not forget that I’ve left behind the vision of shiny roads and instead I’m seeing “messy”. I see the reality of autonomous vehicles (AVs) struggling to cope with roads shared with messy humans, I see systems talking fluently over the data highway only to fail at the user interface, and I understand the implication of the second speaker’s message - “it is easier to predict machine behaviour than it is to predict human behaviour”.
Interestingly, the Digital Greenwich project is testing a fleet of AVs this year – but not on public roads as many of us might imagine from the news stories of Google cars. Rather, the trial will be run in controlled ‘sanitised’ areas. This is, in my view, exactly the right approach. It gives time and opportunity to assess not only the technology and its capabilities (and limits of those capabilities) but also the design of the vehicles themselves – how people interact with and react to them. Do they gain human trust? Can they be trusted? Should they be trusted? Like all technology, if we deploy before we thoroughly test, there will be trouble ahead. I was reminded of a piece I wrote in 2014 about artificial neural networks. If we don’t understand what was going on inside them, does it remain easier to predict than a human? I suspect not. Deep machine learning algorithms are not a set of if-then-else constructs but rather a mix of adaptive neural networks followed by classification algorithms and then decision trees. The neural part, of course, is trained by wrong guesses and modifiers, until a high proportion (but never all) of the guesses are correct. There is significant work going on with non-linear transformations, convolutional and deep neural nets and sensor fusion in order to move towards “true AI” – whatever that might mean. It still sounds non-deterministic and messy to me! So, when messy meets messy, what is the result? I suspect the vision of accident free AV travel on a public highway is a little optimistic but maybe that’s a little too pessimistic? After all, AVs have potential access to much more data than human drivers and can respond much faster to changing conditions. Again, I’ll leave you to decide.
Let’s return for a moment to the enormous amount of data that a smart city environment may have to handle. The current intention is to aggregate and process that data in order to provide multiple benefits; Air quality measurements could inform traffic control (did you know that 5% of all deaths in Bristol are accredited to air pollution?). Population hot spots could divert power and utilities from under-utilised zones. Frequency of automated transport services could be optimised according to actual and predicted demand. Stress signs analysis from health data could adjust the environment etc… but this all takes a huge amount of computing power. One delegate asked “how do we control the impact on the environment from all this cloud computing”. And if you think that’s a minor point of interest – think again.
The image here is of a bitcoin mine, where the crypo-currency hashes are calculated. In one Chinese farm, the power usage is 1.25GW equating to a monthly electricity bill of $80,000. And they need cooling. In contrast, I do not profess to know how much computing power will be required in the smart city but, interestingly, neither did the speakers. Few would dispute that it will be considerable and we need to start thinking now about how the impact can be offset by renewable energy sources.
Who is it For?
Labour MP Andy Burnham gave, in my view, one of the most inspiring talks of the day. Let me start by asking the question that for me, summarised his talk – “What is Technology for?”
That’s a great question, and was answered in another great presentation by Andrew Till of Harman International, in which he proposed that technology is about making what is scarce, abundant. It’s worth pondering that statement for a moment. Isn’t that what engineers do? We try to improve scarcity of food, energy, water, health, safety, leisure, knowledge… and time.
Andy’s talk put people at the centre. He implored the industry to consider moving the technology from products to people. Use the technology of the smart city to create opportunities for young people and SMEs – to create new ways of working and collaboration. This could be where the re-distributed data could make a difference, addressing the increasing scarcity of opportunities for our young people with data hubs that provide the latest and greatest training, teaching and work experiences available. Furthermore, we have never really nailed the remote working thing – think of the benefits for the individual, the transport system, the air quality, and the decentralisation of our cities if we did.
So let’s consider who this technology is for. Let’s also consider the social impact. Will we widen the gap between the rich and the poor, the technology adopters and those who don’t understand it? We were reminded on the day that the technology backbone that can support a smart city might not be there – and this could be true for the majority of the world’s cities.
Let me finish with another quote from Andrew Till regarding the scale of the challenges: "It may be too much to ask those dealing with the present to deal with the future". In this new data age, with changing technical, economic and indeed physical landscapes, there is a need to search out and create new opportunities. But with ever increasing demands on the pace of current development and innovation, this statement could contain a valuable lesson - not only for our industry but for business as a whole.
The benefits and opportunities available around the smart city are plain to see but let’s not follow the hype curve. It’s too important. Let’s have all the debates, explore all the options, test our ideas and then do what engineers (and politicians!) aim to do best – improve people’s lives.
Written by Paul Jennings, Head of Technology