Gathering accurate, anonymous, and detailed data insights is essential in supporting transport networks and infrastructure.
Life in our cities is changing dramatically. To tackle heavy traffic flow during peak hours, help reduce congestion and emissions, and help make informed decisions on future infrastructure, detailed data insights are going a long way in future-proofing our cities.
To learn more about the benefits of video-based traffic sensors in capturing these insights, the Cybernews team reached out to Peter Mildon, COO of VivaCity Labs, the London-based transport technology scaleup.
How did VivaCity come about? What was your journey like since your start in 2015?
I met my two Co-founders at the University of Cambridge, where we were running a project to build a solar-powered car and race it across Australia. This involved not only the engineering and logistical challenges of designing and building the car but also raising funding for the whole project. While it was an amazing project to build something that could go at motorway speeds with the power consumption of a hairdryer, unfortunately, 3 days before the race in 2013 in Australia, we crashed the car and we learned what it meant to fail.
After a couple of years of working in consulting, we re-grouped and decided we wanted to get back into the world of sustainable transport. Looking at the automotive industry back in 2015, however, it was going to be challenging to compete with the likes of Tesla, so we turned the equation on its head and asked ourselves “can we make the road smart instead?” This is where the concept for VivaCity was born.
We initially looked at traffic lights, wanting to focus on how we could prioritize cyclists at key junctions. However, an initial market scan showed a lack of high-quality cyclist sensors that were able to cope with a truly multi-modal road. This is why we wanted to start building a new, AI-powered cyclist sensor.
We grew to about 10 people by the end of 2016, bootstrapping the business off the back of some Government funding and innovative clients. In 2017 we raised our first Seed investment round which enabled us to accelerate growth and the development of our product. We have now grown to 120 people, and have sensors deployed in multiple countries and over 100 UK local authorities.
Can you tell us about what you do? What are the main problems you help solve?
VivaCity’s core product is a video-based traffic sensor. We take data privacy by design very seriously, ensuring that this sensor device only produces anonymous data on traffic movements. This sensor device consists of a camera, a local GPU processor, and some machine learning software that can extract anonymous data from the video feed. The video is then discarded, and the anonymous data is sent to the cloud to a database and API access for our clients.
The sensors not only provide basic classified counts (pedestrians, cyclists, e-scooters, motorbikes, cars, taxis, vans, heavy goods vehicles, and buses) but can provide a richer dataset including assessing turning movements, the proximity of vehicles, and the safety of maneuvers.
This information can then be used by road authorities to better understand how their roads are being used, the investments and schemes they are working on delivering the expected benefits, and assess how the road spaces could be made safer.
Would you like to introduce us to your smart city concept? What are the cities going to look like shortly?
Cities have ever-changing demand for different modes of transport (e.g. e-scooters becoming more prevalent, e-cargo bikes the next thing to take off, and ultimately autonomous vehicles starting to displace driven cars).
The venture-backed mobility providers are disrupting the market, with Uber transforming how we book taxis, and the e-scooter providers almost “dumping” products in cities outside the UK overnight without any regulation.
Local Transport Authorities keep transport systems functioning in this fast-changing world; our job is to give them the tools they need to do that.
VivaCity’s vision for Smart Cities is to empower the Local Transport Authority to:
- Better understand how roads are being used, in real-time
- Be able to react to this data through automated controls (e.g. traffic signals, routing)
- Plan for future schemes and infrastructure
Did the recent global events present any new challenges in your field of work?
We doubled revenue in 2020/1. This was helped along the way by a government initiative to increase cycling creating an appetite in being able to measure the impact of these initiatives.
Alongside this, VivaCity launched a “Social Distancing” module that was able to anonymously measure the distance between pedestrians, providing critical information to policymakers on the public’s reaction to various messages during the various lockdowns and rule formations. It is impossible to say what would have happened in a no-pandemic world, but for us, it was a period of strong growth and market expansion.
On the other hand, global supply chain issues for semiconductors have certainly added a new dimension of stress to running hardware-based SMEs. We have had to increase our stock holding as a company, but have managed to weather the storm and not had any major delays to deliveries due to these challenges. Achieving this has certainly added to the grey hairs growing on my head!
Since transportation infrastructure is one of your main fields of focus, how do you think this sector is going to evolve in the next few years?
Autonomous cars are coming, but not as fast as the market thought they would 5 years ago. I think we are going to see a big increase in the number of trial systems on public roads around the world fairly soon, but we won't be driving alongside autonomous vehicles country-wide for quite some time to come.
Other modes of transport will continue to pop up, and some of these might be drone-based (either flying or on the roads). I expect chaos initially, followed by hopefully some improvement across the board.
Before the pandemic, I used to answer this question by saying that once VR became much more realistic, we might see transport demand reduce, as people began to virtually commute to work, and chose to live where their personal life dictated. This, I predicted, would lead to reduced transport demand.
While you might say that the pandemic tested this theory to the extreme, I think it might have delayed it, as the fatigue from constant video conferencing is going to remain in our collective psyche for some time, even after VR can produce the same quality of human interaction that we can achieve in an office.
What predictions do you have for the future of AI-powered solutions? Do you think this technology is going to become commonplace soon?
We have proven that an AI algorithm can reduce delays on the roads at traffic lights by 23%. This is a complex problem for the AI – compared to the video games that reinforcement learning was originally developed on, it has thousands of independently acting players (road users), and different junctions will interact with each other in complex ways. We have found that AI is a fantastic tool for pushing the boundaries of what was previously thought possible in automating this type of system.
AI has already become commonplace in our day-to-day lives, all of our mobile phones already use it in video processing and voice recognition. The local government market might be a bit slower to adopt the technology, as AI is a black box – it is not easy to know exactly what it is going to do in a specific circumstance, and this leaves some in the industry slightly uncomfortable.
We don’t see this being a blocker to the technologies’ adoption, however, as a set of deterministic rules can sit around the AI, giving it enough freedom to find new ways of optimizing traffic, while offering reassurance that no one will be left with a red light for more than a couple of minutes.
What other aspects of our daily lives do you hope to see improved by innovations in technology?
For me, the biggest challenge facing us is climate change. Technology is not going to be the only solution to this challenge, and while there are plenty of opportunities for AI to be used to help make the world more efficient, we are a very long way off from AI being able to invent new energy sources for us!
Technology other than AI is going to play a big role in helping solve climate change, and the next big innovations the world sees need to be in this area.
In this age of ever-evolving technology, what do you think are the key security measures everyone should implement on their devices?
Good password security should be taught at primary school! Our online identity has become such a big part of our lives, that we get asked for passwords to everything all the time, exposing us to ever more risk. If everyone knew how hackers tried to steal our passwords, and the mechanics of how to stop them, we would all be much safer.
Would you like to share what’s next for VivaCity?
Pushing our AI traffic signal control product out into the market and towards general adoption is our next big priority. This has the potential to save anywhere between 20% and 50% off journey times through congested junctions without having to build a single meter of road.
Beyond the obvious capacity potential of this, it also becomes a tool in combating climate change – as greener modes of transport can be prioritized, and fewer idling vehicles produce a more efficient city.