News | 21 October 2021
TTF Congestion Challenge event

Nearly 100 industry professionals met in York on 21 October for the TTF’s Congestion Challenge Event.  Here’s a summary of proceedings.  A recording of the event will be released in due course thanks to our friends at ITS (UK).

Here’s a summary of the day:

Anthony Ferguson, Deputy Director at the Department for Transport welcomes everyone to the event.  He described congestion as “the biggest problem”.  He says if he gets asked about congestion and what the DfT is doing to solve the problem, many might say they are doing something to fix the problem but the answers are varied – “public transport is the answer” or “active travel is the answer” or congestion’s environmental impact is being tackled by electric vehicles.

“One of the reasons we keep coming round to it, he says, is because it’s so hard to fix,” he said. “One of the exciting things of the last few years is the funding to find new solutions to congestion, and technology is a way to do things differently.  Better information, better tools, better services is used efficiently.  These are the sorts of things the TTF is here to do.

“We haven’t been in lockdown sitting on our hands doing nothing – everyone has been really busy and so we can hear the projects that are starting to show promise that means here-and-now problems can be solved by here-and-now solutions.”

The next speaker is Dave Atkinson, Head of Highways and Transport at City of York.

“York has got a deep and rich history,” he explains.  “Founded by the Romans at the start of the first millennium.  What happened when they arrived is they dropped the scroll with the rules on EV charging, congestion management.  In the Year 900 we didn’t have an e-scooter strategy.  York has a history of innovation and social conscience which you can see from the chocolate industry.

“When we talk about our e-scooter trial with our residents, it has led to either a beaming smile or recoil by people.  But we think we have shifted about 16,000 car journeys onto e-scooter.  That’s a drop in the ocean, yes, but it’s all about small changes and the things that excite me most about the e-scooter trial is that we have introduced a new mode of transport and a shared mobility scheme.  Shared mobility plays a fundamental role in changing transport.

“On EV charging, our hyper hubs have been built and are about to be switched on, and that would effectively give us two top-class ultra rapid and rapid facilities on the outskirts of the city to deliver resident charging and people travelling.  We have solar harvesting and battery storage to deliver sustainable charging.  In the city we also have in plan a third hub in the city centre.  That will go through planning at the beginning of next year.”

We now move onto projects.

Mark Nicholson from Vivacity talks about PROACT talking about how traffic control centres can be proactive, rather than reactive.

“Control Rooms have a range of information sources – CCTV, floating vehicle data and calls from emergency services or the public, and the output are signal timings, messaging and other solutions.  This is time-pressured for people managing the traffic, dealing with inputs and outputs in a very short period of time.  That’s a reactive control room.  This is true across a wide range of cities.

“Here’s an example in Bournemouth.  Outside the hospital there is an issue with congestion – the goal of this project was to find a solution and offer a broader solution to other authorities across the country.

“Our solution involved better live data, better diagnosis of the problems and how to solve them.  We need to make it easier for them to deploy a solution in the real world.

“The solution means we use better data sets – computer vision powered sensors to understand what is happening on the road – we get movement, turns, speed, occupancy.  The basis of that project is using all this data and feeding it into the control room to understand what is going on.

“How do you define a congestion problem from that data?  How do you draw the line on what is congestion.  In a manual way of doing it, you can look at CCTV and get a “finger in the air” solution.

“We have defined it in three areas – the occupancy, speed and journey time.  When these are combined you can define a really granular definition of congestion.

“Understanding congestion is the first step – alerting the Control Centre is the next. We use Dynamic Historic Averages – we use a rolling average of the past few weeks and compare to the present situation and see if what is happening now is abnormal.  The next step is to use Machine Learning to have an algorithm to tell me the abnormal events.  You need to have many previous deployments to teach them.

“These alerts then were sent to operatives alerting them to any abnormalities.  These operatives can then see images to see if there is a traffic jam, with a broad data set of what is happening city-wide.

“Implementing the fix involved existing solutions with VMS, signals etc.  The next stage is to change the signal timings automatically and we are working towards that because you need real confidence in the system to implement that.”

“Regarding Covid, we have learnt to manage traffic from home, so we virtualised the control centre which is an interesting transition.

“The project outcomes were indeed live data, better diagnoses and solving the problem.  Within the project we proved that this was making a genuine difference on how authorities can manage these.

“How do you feed this into signal control?  Currently we have congestion alerts going into UTMC and we use a pre-defined control plan.  These need effort and the control plans are never perfectly aligned to the problem.  So we are using AI to react and adapt much more dynamically.

“Live data uses occupancy, speed and journey times into microsimulation – we run thousands of simulations to train our AI into how to manage signal timings – given a particular scenario how do I manage the traffic?  We have done real world AI solutions in Manchester and Cambridgeshire.  We have learned we can reduce the average journey times by 23%.  We were that much better than Scoot.”

Dave Williams of Immense talks about the SATO project.

“The challenge is a real-time data tool capable of forecasting traffic impacts as a result of changing conditions on the network, and therefore to aid in active congestion management.  The tool has been developed through close engagement with a range of local authorities using transport modelling.

“We focussed on user needs, developed a web-based user module.  We have tried to solve – “user needs analysis” – using user experience design.  Our products need to be useful, easy to use and add value.  I think we’ve all used software that doesn’t do that, so we made sure we did.

“The problem is competing demands for roadspace, a need to manage changing circumstances, proactive rather than reactive, an acknowledgement that impacts are wider than a localised event and when experienced personnel change roles, knowledge is lost.  So our tooling needs to help us deal with this.

“Another crucial thing is “who is trying to solve it?”  What we found is that traffic managers, front line staff, control room operators and signal engineers need to be experts in their job, it’s not realistic to be expert in all facets of data and simulation.  This is all with shrinking budgets, lack of integration, legacy tools, and a desire to do more with less.

“We have developed a web-based simulation tool that enables the impact of changing conditions to be forecast, leveraging new and better data sources.  We design with the users throughout the process – the challenge, problem definition and outcome.  We have worked not only with Oxfordshire but a range of authorities to work out the jobs to do.  What problems are they trying to solve?  What workflows do they want and what results to they require?

“We use simulation when we want to test different things – it is almost always cheaper, quicker, easier and safer to test things in a simulation rather than in the real world.  This tool makes it easier to test scenarios.  The basic problem we are dealing with is route choice.  Crucially, we need to ask “what if” and what are the impacts.

“Real time data helps us make our simulation better.  The data we rely on has changed a lot over the past 20 years.  The old ways we understood a “typical” but not variations.  The data we have now is so good, it’s unacceptable to use a “typical” day now – typical days never happen.  Traditional models relied on typical days (which are OK for a business case) don’t work for managing a road network.  So we tailor the model more closely to what is happening now on the network.

“Key enablers is a solution has to be useful – does it make a demonstrable difference to the user’s job?  It needs to be economical, productised data is key, but signal data remains an issue, and there need to be better standards around signal data – we find some of it is still written on timing sheets and doesn’t make it easy for us to deploy elsewhere. That is why we are working with INRIX on data.”

Next up is Bradley Taylor from INRIX and Andy Graham from White Willow Consulting talking about the PATH project.

Andy Graham: “All the projects we’re talking about were funded by Small Business Research Initiatives, to come up with things that local authorities could actually use.  Our challenge was about signals – including temporary and signalised junctions – and where there are junctions that should have signals.  So we helped understand how to prioritise spend.  We used INRIX’s anonymised floating vehicle data.  What we looked at was individual vehicles – we do analysis on that to understand delays.  We don’t just have cars but thanks to BODS we have buses too.  We also have connectivity.

“We found very early on that one size does not fit all.  We had a key user in the City of York but had other local authorities for testing: Birmingham, Bedford, Dorset, Portsmouth, Milton Keynes, Wolverhampton, Oxfordshire, Bristol, Liverpool, Hertfordshire and Coventry.

“The main themes from user research included numerous systems with limited methods to compare historic traffic data, a difficulty to determine if an intervention has made a difference, local authorities want to be proactive rather than reactive, and you need to be able to zoom in at a junction and see if the impact widens out.

“There are three themes to it – data, metrics and tools.”

Bradley Taylor: “The theme is to save time – it’s about saving the engineers the time of understanding the traffic.  We identify corridors, produced a heat map understanding journey time and speeds through a set of junctions, a “curtain bead diagram” seeing individual trips and then giving a junction view.

“We have delivered a prototype and we are now trying to make it scalable and to integrate this into what we call INRIX Signal Analytics.  The main user cases from PATH was having a quick view of the network – it’s now about how we can bring the data of our whole road network.  We have created a dashboard view We can then compare day-by-day how things are changing and which junctions are performing well and which are not doing so well.

“We can track a vehicle as it approaches a junction, we can see its associated speed and reference that against the free flow speed and then calculate any delay.  This allows us to identify split failures, where people have to wait for more than one cycle.

“We are able to provide detailed information related to movements through a junction.  We can see over time of day how many people are getting through a junction on green or not, and you can see how timing plans change and how many people get through on green, and that can identify where you might wish to refocus your timings.

“The premise of this project is to compare – have I actually improved anything?  It takes an understanding of your network, identify issues, make changes and see what improvements there have been.

“We have aligned with the Traffic Signals Maintenance Fund, which required evidence against seven themes.  We think PATH can help surrounding prioritising upgrades, reducing emissions and providing a monitoring mechanism.

“In summary, we can provide a daily look at the junction performance on your network.  Viewing junction performance metrics, drill down, identify junction performance, see level of service values network-wide, run reports and reduce the amount of time you have to spend analysing your network.”

LUNCH BREAK – EVENT LUNCH SPONSORED BY INRIX.

Tessa Hayman from Aimsun is next talking about the NEVFMA project to manage traffic-related emissions in the Oxford Area.  NEVFMA stands for Network Emissions/Vehicle Flow Management Adjustment.  This fuses real time data with transport modelling and emissions.

“The objective was to mitigate the air quality problems due to road transport across Oxfordshire, utilising existing council infrastructure which would compare different traffic management options, including do nothing, and deliver the best mitigation strategies.

“The model predicts in 15, 30, 45 and 60 minutes into the future, testing and comparing response plans in real times with a real time emissions model.  The product would monitor, predict using historic and real-time data for traffic patterns and profiles and, if there is a problem now or a problem predicted in the near future, offer decision support.

“The Aimsun Next traffic model was based on ten different profile types, using ITS detectors and air quality monitoring to deliver that support.

“We noticed every single day is different and things fluctuate.  So, for example, we don’t have a “normal Monday” – we analyse the historic data, identify common patterns and then in real time we identify the current pattern, provide analytical and simulated predictions and current data becomes historic, so we are constantly learning and fine-tuning the model.

“Real-time simulation offers wider coverage, understanding of non-recurrent events and alows you to emulate signals and test response plans in real time and make sure they are fit for purpose.

“One of our use cases was not just in real-time unplanned events but also roadworks.  We have analysed the effects of a full closure of the Kennington Bridge across South Oxford.  Using modelling during planned works can test the effect of traffic management on route choice, traffic patterns, the effect of plans on air quality and rapidly test response plans for unplanned events.

“Looking at another example on the Botley Road into Oxford, we found the modelled NOx exceeded legal thresholds on 57% of days and response plans modelled reduced emissions on 60% of the time with an average improvement of 5% of the time.

“In summary – typical days aren’t typical, the system learns and adjusts the longer it is deployed real time systems can aid for planned and unplanned events.  Thanks to Innovate UK and National Highways for funding the project and Oxfordshire for letting the county’s roads to be used for the project.”

Tim Rivett from RTIG is next talking about the Bus Open Data Service.

“BODS holds routes and timetables, location data in real time and fares, with an encouragement to supply occupancy data as well.  We have had open data for schedules for 20 years but the Bus Service Act, which covered a number of things including BODS.

“For timetable data, from the beginning of this month all the data is in a consistent format, we have 20,000 vehicles supplying data at the moment, which is a very good proportion of the vehicles on the roads, while we are in the early stages of collecting fare data.  They haven’t got long because they’re legally required to supply it by the end of the year.

“What I want to talk about is how you might want to use this – when you get high quality data in high volume is that you can use it to analyse.  The Analyse Bus Open Data Service offers a dashboard showing, for example, on-time performance, which plots all the data and shows where the bus should be and where it is, which helps understand what is going on on your network.  This is useful to see where the bus might lose time, so there are things you can do physically or change traffic signals, for example.

“This is available to every authority in the country.  Hopefully this is a bit of a game changer.  It gives historic data for congestion data but also the here-and-now which you could use to inform how you could provide traffic light priority for buses.  If there is one take-away from today, you should be looking for colleagues who are working on bus service improvement plans.

BODS provides a single data source in a single place so is easier to integrate and easier for providers to share their data, which makes life easier.  We will soon be looking out for pilot areas to do some work using BODS data on bus priority.  One of the challenges at the moment is that if you are looking for traffic light priority, you really need to know where buses are, so we need something a bit faster than the average 17-second update cycle – if we can push that we can pull together a project to see what we can do with that – come and talk to us!”

White Willow’s Andy Graham then chaired a session on in-vehicle information projects.

James Guilliatt talked about the Smart Transport Evolution Project STEP in York.

“STEP prepares York for CAVs and enables next-gen traffic management and opens up data.  Have brougth together data from a range of sources and works with CA Traffic, PTV Group, White Willow, FLIR, Dynniq, National Highways, Interoperate and Alchera Technologies.

“Our data uses traffic flow counts – we have 100 traffic flow counters in the city, half of them are delivering back with less than one minute latency.  This is fed into an Optima real-time model. Then there is signal data, speed data and highway condition information for roadworks data.  The benefits have been to give more knowledge to control room operators and external systems via an API.  We can then make automated decisions based on KPIs.

“Another part of the project is GLOSA – green light optimal speed advice service, which displays the live signal heads in an app and gives them an advised speed to drive to arrive at the lights on green.  In York this is deployed via Dynniq’s car flow, it’s public-facing, delivered over cellular communications.  We want people to be able to use this service regardless of wealth – it isn’t just the fancy cars that can use it, anyone with a smart phone – which most people have – can use the service.  We want to see how the public uses this service, and we’ll then share this with the DfT and other road authorities.

“This is a three-year trial so we can see how people use it, and if the two corridors we have now are a success we may be adding more, but there are no plans for the city centre.”

Paul Rose of Amey then gave more details about Glosa, off-slip trials with National Highways, Transport for Greater Manchester and virtual VMS in trials with Oxfordshire and Bristol.  “This saw a reduction of 27% average CO2 emissions by HGVs, that is 17% average reduction in NOx.  This works out at 12.5p per vehicle.

“In Manchester, using Open Data, delivering to different apps, using legacy infrastructure to deliver Glosa, with the virtual VMS project looking at what you need to do to legacy systems to get VMS data out of a system to get the message into a vehicle.  This is about getting a standard to get that data out, which is based on C-Roads.”

George Brown from KL Systems then demonstrated in-vehicle VMS with a number of videos, followed by Eloy’s Damian Horton using an in-car app using Apple Car Play with more demonstrations and Zircon talking about the M5 Junction 23 Off-slip Glosa project to influence drivers on a MOVA-controlled junction using a machine-learning algorithm to predict the next decision MOVA is likely to take.

Dave Gabbitas of Yunex Traffic gave an update about SPAT/IVI, including Project Synergy to deploy Glosa at a site near Manchester Airport.  This included four IVI message types presented at the junction. The displays show signal states from a large distance with a 3D map, but as you get closer it shows the signal states of the signal groups ahead.  Also working with MIRA on two SPaT installations, the autonomous bus service across the Forth and upgrading sites in Portsmouth.

Finally Caura explained how he has built a payment platform for your car, using just one product based on your registration number.  The idea is to take the pain away from owning a vehicle so you can pay for insurance, parking, tolling etc.  These are things that are important to drivers.  To make these payments on Caura’s platform you make one tap and it’s done.  We make money on cross-selling and upselling insurance products, MOTs etc.  The whole point of this is to integrate these services into the vehicle companion apps, and we are doing a trial with Jaguar Land Rover.

Daniel Hobbs then talked briefly about the Manual for Smart Streets which he is working on – the intent is to bring all that has been heard from today into one coherent set of systems, while Ben Turner from Innovate UK who described it as an “exciting time” for Innovate with their 2050 Transport Vision to get the conversation going about innovation in the UK, creating conversation for the industry to give feedback.  This year Innovate UK has invested £130 million in transport this year which are being showcased.

“Before I worked at the DfT, I worked at City of York so it is great to see so many successful projects here in the city,” said Darren Capes from the Department for Transport who manages the TTF.  “What we have heard about today is showing real products that the DfT can use to solve problems.  The link we have with RTIG is important to the TTF so we can work with different sectors to bring solutions together. The In-Vehicle Information presentations was mind-bending.  We’ll have to have a whole event on this – it’s relevant to everyone on creating services people want.  That is the future.”

(Picture shows Chairs of the event, Anthony Ferguson and Steve Gooding)