Introduction About us ITS Features ITS News Links Contact us
 
Class of 2005 - Pre-judging the issue.
The following article was written for Diamond Consulting Services, the developers of the IDRIS product.
10/12/2004
Growing concern over increased levels of congestion at tolling plazas and their inherent inefficiency is leading to a radical re-think on ways of improving the through-flow of traffic. In the US this has led at least one major tolling operator to look at the option of pre-class tolling despite the considerable difficulties that such a course represents.

But while the issues require extremely careful consideration but there is no reason to believe that the idea should not deliver the benefits expected of it.. Technology has already proved itself more than capable of dealing with highly complex traffic situations, including those relating to ETC and automatic tolling environments, and can handle the complex tasks involved although these are systems issues which require "joined up thinking".

The technology cannot however act in isolation and because of the inherent problems, it has only been deployed in a handful of sites. The difficulty is rooted in two quite distinct areas. The first of these is the environment in which pre-class tolling is expected to work - heavily congested stop/go traffic conditions that require the technology to classify, separate and track each vehicle as it moves through the system. The second is automation. There is no room for error in a system which offers little or no prospect of immediate redress in the event of failure.

"The technology needed to work in the pre-class environment is hugely complex," said Andy Lees, co-developer of the Idris technology and systems engineer with UK-based Diamond Consulting Services. The dilemma is that without pre-classification, automatic toll collection using coin machines does not offer the ability to enforce a variable tolling regime. Tolling operators have found themselves being torn between the drive to automate i.e. reducing the number of manual lanes and the labour costs associated with them, and the attractive economics of variable toll schedules. Offering all lanes to all classes of vehicle is clearly preferable to the tension-ridden process undergone by drivers anxious not to be caught in the wrong lane. A coin-only automatic lane allows approximately double the throughput of vehicles but only allows one fee unless the vehicles are pre-classified prior to reaching the payment point.

"The only other alternative," said Andy Lees, "is to dedicate certain lanes to specific vehicle types, which in itself leads to inefficiency and an increased likelihood of unacceptable delays to drivers."

What it amounts to is that in the absence of an efficient system allowing all classifications of vehicle to use all the available lanes, operators face a future blighted by the prospect of continued inefficiency and its corollary, falling customer satisfaction. Yet there is no question but that the routine use of technology has made significant improvements to the rate at which traffic is processed through toll plazas. Rising to the challenge presented by the rapid increase in traffic volumes is at the top of the agenda for many toll authorities. In situations where a large proportion of the traffic belongs to occasional users, such as tourists who make up a substantial proportion of toll customers in states such as Florida, the "holy grail" of non-stop Open Road Tolling (ORT) does not solve the entire problem. Unfamiliar with the tolling regime and lacking the necessary equipment which would have enabled their use of ETC (Electronic Toll Collection) lanes, visitors are forced to make use of the slow manually operated option.

"If you want to make full use of all your lanes," said Bob Lees, brother to Andy and an electrical engineer with over 20 years experience in the traffic business, "then pre-class tolling is essential. It is crucial that you know the classification of the next vehicle in the queue so that the correct toll fee is taken. Pre-class systems allow any vehicle to use any lane."

With full utilization come the benefits of less congestion, better customer relations, and a more cost effective operation. It is also far quicker. A fully automatic barrier payment system will deal with roughly double the number of vehicles passing through a manual system. By contrast a non stop, Open Road (ORT) system can deal with ten times the volume of a manual system but such systems rely on regular users with the necessary transponder.

In October this year, the Florida DOT issued an RFI (Request for Information) specifically aimed at the provision of pre-class AVC (Automatic Vehicle Classification) throughout the State's 750 tolling lanes. In Florida's case the clearly expressed preference is for an in ground solution.

"The technology required to operate a pre-class system is exactly the same as that needed in a post-class array and there is no reason, other than cost, why the same arrays should not be used for both functions" explained Bob Lees. "The reluctance to use above-ground systems in the pre-class environment is based on the technological complexities and the extra costs involved. The equipment is always vulnerable to accidental damage from passing traffic and has to be protected by expensive barriers. This is especially true of the classification arrays which need to be at least 25 metres in front of the payment booth so as to accommodate large trucks."

Whatever the reason, it is to the in-pavement solutions, including piezos, treadles and inductive loops that authorities like FDOT are now looking for their pre-class systems. Are they right to do so or is the current development of tolling technology simply not yet robust enough to withstand the pressures of pre-class tolling? Certainly there are those within the industry who are prone to a shake of the head when any mention is made of the idea. Yet what such an attitude fails to take account of is the very real and substantial advances that have been made, particularly in the field of vehicle detection technology using inductive loops.

The dismal performance of the loops of a decade ago, when error rates in the order of 1:100 were the accepted norm, where the analysis of data was barely sufficient to satisfy even a statistical package and the range of applications was so small as to be almost invisible, can safely be consigned to the dustbin of history. Over the last seven or eight years, Diamond Consulting Services, a small research and development company headquartered a few miles outside London has developed its patented Idris product to the point where count (separation) error rates are now measured as better than 1:10,000 and classification accuracy levels are said to be in excess of 99.8%. Operating under all weather conditions and virtually maintenance free, the software-driven package is now used in tolling operations throughout Europe and the USA, separating and classifying vehicles as they pass through the system.

There is little doubt that an AVC system like Idris would equip itself well in the pre-class tolling environment. Its strength lies in the real-time analysis of the signal generated as a vehicle passes over the loop array. This analysis of the vehicle's signature produces the "Per Vehicle Record" or PVR allowing Idris to accurately track vehicles through the system. The available data includes the vehicle's length, number of axles, type (profile class) and speed of travel. It will, says Andy Lees, routinely identify the difference between tailgating vehicles and those joined as trailers and will readily present the correct classification. It is wholly unaffected by fog, ice, snow or other weather conditions and can normally be expected to last for the lifetime of the pavement without maintenance. Yet despite the levels of accuracy and reliability claimed for the product, Bob Lees is the first to admit that unless some kind of supporting strategy is put in place, problems could continue to surface in the pre-class tolling environment, due to the high level of congestion.

"It's important to realise," said Lees, "that no system is perfect and there has to be some means by which the technology can correct itself. One way of achieving this is to have a secondary AVC between the classification array and the pay booth which will serve to check and, if necessary, assist in the correction of the original classification. Idris is designed in such a way that errors can largely be eliminated through a self-correcting process."

Contributing to the complexity of the task of pre-class technology is the distance from the pay booth at which the classification array must be placed. There is little point in having it so close to the pay booth that a long vehicle fails to clear it, but the further the distance, the greater the number of smaller vehicles that can (and invariably will) fit into the space.

"The distance between the classification array and the pay booth," said Andy Lees, "needs to be in the order of 25 metres so as to accommodate the longest vehicle likely to use the plaza. The trouble is that that area is large enough to squeeze in four or five cars, each one of which needs to be classified and tracked as it approaches the payment point."

Slow moving as it approaches the booths, the traffic will naturally tend to bunch-up and provide a further headache for the technology. Fundamental to the success of the venture is an ability to separate one vehicle from another, to tell the difference between two vehicles and a lorry with a trailer. A failure to do so will, at the very least, lead the system to expect one more (or one less) vehicle than is the case causing difficulties for following traffic. Lane barriers are an essential adjunct in all systems but the cost of providing the substantial structures needed to protect above ground systems from damage would be considerable. True, barriers would anyway be needed to funnel approaching traffic into single lanes but these could, if not required for protection, be of a lightweight and therefore inexpensive design.

One might also add another cost to the ongoing provision of conventional classification technology (light curtain, treadles) that does not apply to in-pavement systems using only inductive loops. Exposed to the elements, above ground sensors require frequent attention. At least one major US tolling authority has its toll collectors routinely cleaning light curtains at the start of each shift.

In addition, treadles also have an inherent weakness. Containing moving parts, they will fail after a finite number of operations - usually around one million activations - requiring replacement about once a year with performance tending to degrade towards the end of a sensor's life. Such degradation does not occur at a predictable rate which means that errors are not always picked up at the point they start affecting performance.

Given the will and the right technology there is, according to Bob Lees, no reason why pre-class tolling should not work and work well. The economic, environmental and customer relations benefits that flow from such a system make it worth the closest of examinations. Technologies like Idris have already proved themselves more than capable of dealing with highly complex traffic situations, including those relating to ETC and automatic tolling environments. But choosing the technology is a matter entirely within the gift of the service provider and all that this article has sought to do is point out some of the advantages and disadvantages of the available technologies. More than this it cannot do.

Further information, contact:
Teri England, Diamond Consulting Services, Buckinghamshire, England. Tel: #44 (0) 1296 747667, E-mail: teri@diamond.demon.co.uk
 
back to top
return to features