In detection and monitoring precise identification of both the subject and its intent is crucial to success
As the technologies and stakeholders in traffic management evolve, what can we expect to see happening in the coming years?
For many, the conversation of the moment is just how, and how far, the newer technologies and services provided principally by the private sector should be allowed to intrude into the realms of traffic management.Traditionally, traffic management has been the preserve of the public sector but, statutory obligations aside, ‘tradition’ was in most respects driven by having the budgets for and access to the types of technologies which were up to the task. That no longer holds true as solutions such as Floating Vehicle Data (FVD) and cellular tracking extend their numbers and geographical reach. A growing number of private-sector organisations are now offering, or preparing to offer, traffic management – as distinct from traffic information – data services.
Perhaps inevitably, strident claims are made regarding the benefits of emerging technologies – claims which in many instances pander to the aspirations of potential customers rather than reality; the demise of the inductive loop has been predicted for decades, for instance, and yet it continues on evolving and providing sterling service. But one thing is certain: both the types and numbers of sensors we see are going to continue to increase for the foreseeable future.
A question of application
There are distinctions to be made whenlooking to compare the relative merits oftraditional and emergent solutions fortraffic data collection, says Dieter Cosaert,“FVD is very useful for applications such as predicting travel times,” he continues. “It’s much less effective at supplying occupancy and vehicle count data – something which more traditional cameras and loops do very accurately. All currently deployed traffic management systems need more than just vehicle speeds, so FVD remains an additional tool, not a replacement.”
Data extraction is also more complicated with FVD in some circumstances. Cosaert notes that the deployment environment plays a large part here.
“For providing travel times on strategic roads, such as highways, FVD is fine. Move into the urban environment, however, and things change. How do you distinguish between a parked car and one which is sitting in traffic, for instance? What happens when a vehicle is slowing because the driver is looking for somewhere to park? Such things will have an effect on perceptions of what is actually going on. Traditional sensors remain much better suited to the more complicated tasks.”
The same observations apply to the tracking of cell or smart phones: “Again, how do you distinguish whether they are inside or outside a vehicle? How do you know whether, say, a stationary group of 25 cell phones is a bus or pedestrians standing at an intersection? You can extrapolate data but still not as accurately as when using more traditional sensors. There just isn’t the accuracy down to the individual cell in a cellular network.”
For those reasons, he says, traditional sensors will continue to maintain their place in the market. A further limitation as regards FVD is initial cost.
“To set up an FVD network from scratch would be hugely expensive. A way round that, to an extent, is for road authorities and jurisdictions to buy into some of the FVD-related work currently being carried out by universities and other research organisations. The pace of research will perhaps be a driver of adoption but I still think we’re a long way from commercial viability in many respects. Even then, I’m sure more traditional systems will be required and specified for critical points on the network. Where I do see the two coming together in the future is in validation of each others’ outputs. But again, that’s very much a complementing rather than a replacement activity.”
New traditions
Econolite, another long-term player in the ITS sector, is positioning itself to make inroads with newer technologies.“Vehicle and smart phone interaction with the intersection is on the horizon and once it arrives it’ll be here to stay,” says the company’s Frank Provenzano. “Similar technology convergence will be demonstrated at this year’s
“A year from now you won’t be able to buy a phone that isn’t a smartphone but how that phone interacts with the vehicle is already evolving. Ford and Microsoft, with SYNC, have proven that the technology is both affordable and sells more vehicles. Android and Google have allowed interaction to go one step further. Consumers are driving developments in many respects. People want quality of life improvements. They want in-vehicle navigation but not necessarily discrete navigation systems.
“In the US
In an urban setting, assimilating information from mobile and cellular devices is especially difficult and Econolite’s choice of TrafficCast as a partner reflects the latter’s algorithm development experience and expertise, Provenzano continues.
“The newer technology developments are all about convergence, not replacement. We have to live with development and look at technology in that way. I think that rather than feel threatened by the consumer electronics sector, the ITS industry needs to embrace it and use it to its advantage. This new technology works very well so we have to think about how Bluetooth travel time data sourcing can affect signal timing changes, for instance. Convergence makes better sense – hence Econolite’s modular approach to design. For example, even though our Autoscope video detection system is 20 years old this year, because of converging technologies we see it as having another life, rather than passing.
“Privacy is an issue we can’t ignore and I think there’s responsibility here on all sides – corporate, governmental and individual. We’re going to have to look very carefully at how services are administered, but there are already plenty of examples, from online banking to individuals’ cell phone contracts, of how privacy is handled well. What we shouldn’t let happen is for small single-interest pressure groups to dictate our whole industry’s direction.”
Differences of approach
Nick Cohn, Business Development Manager, TomTom Licensing Business Unit notes that the very different natures of the technologies used by the public and private sectors for detection and monitoring result in great differences between how travellers access and use the information they provide.“Historically, the main drivers of development in detection and classification were in the public sector because of the associated planning and investment decisions, as well as there being a fundamental responsibility among governmental bodies to collect and disseminate data.
“The private sector invests in technology primarily for commercial applications – typically consumer and fleet markets, but paying customers. So, rather than providing an alternative to the public sector, commercial organisations are often addressing another market entirely.
“The main advantage of fixed systems, if they’re properly maintained, is that they provide reasonably accurate volume data. Maintenance is a problem, though, especially for loops, as there’s less and less inclination to close lanes to allow work to take place.
“Another issue is that government agencies like to retain complete ownership and control of their technologies even when contracting out operations. That’s an incredibly costly way of doing things and it doesn’t provide the coverage and efficiencies that they’d like. The concern over loss of control is understandable but with funding slashed the possibility of extending geographic coverage has been even further restricted.
“The newer technologies developed by the private sector can provide speed data but, currently, cannot provide real-time volumes without some form of data adjustment. That’s changing, however, as is traffic management itself: traditional approaches rely on speed and count data but more recent approaches don’t need counts, just a good sample size, location referencing and the appropriate filtering.
“Particularly in urban areas, if you want to use speed data obtained from cellular telecommunications networks you need to apply filters as you don’t know whether the phones being tracked are in-vehicle, being carried by a pedestrian or on a train. It’s not an easy thing to do. But there is a greater and greater focus on GPS, in all types of devices, and with clear identification applications.
“That concept ties in with a number of developments. Governments are looking for alternative financing methods based on what people actually do in terms of transportation habits. There will be less of a direct emphasis on fuel taxation. At the same time, consumers are using systems which offer added value; typically, up-todate maps in conjunction with GPS-based navigation. We’re seeing a lot of market saturation in that respect.” Questions remain over privacy but Cohn sees a lot of these being addressed by who is collecting and using data, and how.
“If you review news reports from the past few years there are more and more cases of data being leaked from or getting lost by the public sector, and I think that individuals’ trust of the public sector is worsening as a result. I do think that the private sector has a better track record overall; it loses money if it doesn’t safeguard individuals’ details whereas it’s not clear what the sanctions are for governmental organisations which fail to do so. We at
“To address the privacy issue, TomTom uses opt-in systems like those used by social media. Our opt-in levels have always been high partly because we don’t ask for personal information but there’s also been a big shift in the willingness and activity levels of people when it comes to telling the world where they are and what they’re doing. That potential can be tapped into for traffic management purposes.
“If you look at developments in traffic information in the last five years, it’s clear that governments aren’t in control any more. There are lots of private-sector suppliers out there, and there are already feedback loops involving car-to-car communication – TomTom’s connected services are an example. These are causing drivers to make decisions based on information that the public sector doesn’t currently have.
“Connected vehicles are going to play an ever-larger role in traffic management. It’s still a real challenge for governments to work out how to tap into that.
Within the current regulatory environment, it’s going to be hard for them to complete the circle between cars and traffic management centres but with little or no funding to expand their own services, public-sector organisations need to look again at how they do things. These new services can be licensed but the public sector doesn’t always have the means to do that, even though these services could end up saving the public sector money. It’s a situation which I think will become magnified in the coming years.”
A presumption of change
Ted Trepanier, Executive Director for Public Sector with traffic information providerWe are, he feels, seeing rapid change: “In just six years, Inrix has gone from being a provider of traffic information across North America to being an international provider of traffic data and analytics in 30 countries. That includes most of the EU and we’re expanding into Asia, and China in particular.”
Inrix processes data from GPS-equipped vehicles and devices to produce historical, real-time and predictive information services for highways and interstates as well as arterials and city streets. The proliferation of internet-connected, GPSenabled consumer devices, whether mobile phones or actual vehicles, is making crowd-sourced traffic information more readily available, Trepanier says.
“There are significant vertical markets for what we do. The mobile and direct-tovehicle automotive sectors are proving the quickest to adopt our services. While the public sector is well established, growth is not quite so fast.”
Looking at things internationally, there are some notable variations: “Levels of roadway instrumentation are a factor which affects growth. From a maturity perspective, the UK and other European countries are probably leading the way in the public sector, with the US close behind. Then there are some ‘second-tier’ European and Asian countries.
Fixed vision
The assumption that developments in the consumer electronics sector will always benefit the ITS industry doesn’t necessarily hold water, according to Bob Germain, Siqura’s Product Line Manager for Cameras. At least in the short term the move towards smaller CCD and CMOS chipsets could have a negative effect on cameras’ performance, he says.
“With 1/2in chipsets giving way to 1/3 and 1/4in chipsets or even smaller in consumer devices, there’s a real danger that even the biggest manufacturers will wake up one day and decide for themselves that there’s just not enough money to be made by continuing production of the older, larger products.”
The problem lies in the fact that as chipsets get smaller so does pixel size, and smaller pixels absorb less light. For megapixel/High-Definition (HD) cameras to carry on achieving the same definition as with the older chipsets, the only resort at present is to move to the high-end chipsets currently used for machine vision.
“Although machine vision manufacturers will tell you that the costs of their products have come down, there is still a big price differential between what the niche machine vision and larger CCTV markets will accept,” Germain continues. “Machine vision solutions also need larger format lenses, which add further to cost.
“1/2in chipsets remain available but there are bigger questions to be addressed: are niche markets enough to sustain the development of products which take us in performance terms beyond where we are now? As the market moves to HD are users going to accept performance as is, or demand the same performance levels they have with standard definition but with the higher resolution? If it’s the latter we have to think about what the price of the supporting chipsets might be, because there’s little we can do at the back end to offset performance limitations at the front end.
“There are precedents in the cost debate. For standard megapixel cameras, lenses have always been an issue – only when lens manufacturers saw a market did they start to create products. We’re two to four years into the megapixel ‘craze’ but lens manufacturers have only just caught up and closed the price gap. Even so, there remains a premium over Standard Definition lenses.
“At the moment, no-one’s investing in megapixel CCDs. The emphasis is on CMOS sensors. We are seeing improvements there but we’ve not yet seen them achieve the same performance as CCD chipsets. I’m hopeful we’ll get there; we’ve seen great improvements in the move from first- to second-generation CMOS solutions for PTZ camera blocks and in traffic situations with adequate street lighting I’m content that the new blocks will perform well enough. But CMOS isn’t quite there yet in poor light conditions. “We’ve just had our first breakthrough outside of our work with the I-95 Coalition in the US; in the UK, Inrix is teaming with
“On the research side, we’re going to be advancing predictive capabilities to hours and even days ahead. In the case of a major incident, for example, that would include giving an accurate time for conditions to return to normal – whatever ‘normal’ is for that time of day, be it offpeak free-flow or peak congestion.”
This will be some of the most advanced work of its type in the world and was a big consideration in winning the contract, according to Trepanier: “The UK has had a policy of pushing this sort of work into the private sector for some time. We’ll also be doing more standard data integration which though advanced is less unusual. We spool up from early in 2012.”
With more data sources comes a need for more sophisticated management. Trepanier foresees a situation where roadway sensors decrease in number but do not disappear altogether. GPS-based sensors operated by the private sector will continue to increase in number, contributing to increased accuracy from crowd-sourced approaches.
This will also reduce costs for the public sector, as transportation agencies’ build-out needs for physical infrastructure will be reduced; in the near term, greater blending will result greater redundancy and a continuous stream of reliable data, he says.
Going from ‘data’ to ‘information’ means simplifying visualisation. From a public agency point of view, that includes how to present travel times and some very advanced work has been going on at the University of Maryland involving both engineers and graphic artists. In-vehicle, the emphasis is on the safe delivery of content which is relevant.
“The next step is making it bespoke – ‘Don’t tell me what’s normal, tell me what’s different’ – and boiling it down into one piece of information – ‘There’s congestion two miles ahead of you’. Graphical displays are available today. Voice driven interfaces and alerts are still relatively young and need time to mature. Systems such as
