INTERVIEW:The Case for Active e-toll systems

Daryl Fleming, 43, Atlanta-based professional engineer and musician of Canadian extraction, has been in the thick of North American electronic toll design, installation and operations since its beginning in the mid-1980s. With HNTB he was involved in early designs for tolling on the Transportation Corridor Agencies (TCAs) in Orange County CA 1988 to 1991 which has an MFS/Texas Instruments system RF ETC system deployed. At Hughes Transportation Management Systems 1991 to 1992, Fleming participated in development of the Hughes TDMA RF ETC system and initial efforts towards standardization of the vehicle-roadside communications protocol. At Lockheed IMS 1992 to 1995 he was project manager for implementation of the Amtech tolling system on Georgia 400, the Atlanta area toll road, and also worked on the EZ-Pass on the New York State Thruway, and on preliminary work towards the border crossing demonstration program in Nogales where heavy vehicle crossings were automated. Since 1995 he has worked as a consultant to Hughes and Bell Canada on H-407 in Toronto, and with the Delaware DOT on their e-toll implementations of Mark IV equipment on I-95 and DE-1. This work has provided him with experience using all the major backscatter and active radio frequency electronic tolling systems.

Q: How do you start to design an e-toll collection (ETC) system?

A: You have to start with financial accountability. An ETC system is first, and foremost, a financial transaction processing system. It is like running a bank. You have to think security and excellent records. Therefore, financial integrity has to be an integral part of the ETC system requirements. This involves a system architecture that establishes an auditable data stream throughout the application, as well as an integrated, functional collections and enforcement capability. This requires an RF (radio frequency) communications link, or some other wireless link that meets operating needs, in the environment it is deployed, be it a retrofit to an existing plaza or open road. Proper billing requires highly reliable communications through the wireless link at anticipated vehicle speeds as well as the ability to locate which vehicle the transponder is on in a reliable fashion.

Q: The more I talk to managers of toll facilities the more I hear a worry that the e-toll systems are not as reliable as they should be. They say there is a disturbing level of misreads, and no-reads. And this is happening at existing toll plazas with constrained lanes and low speeds. I have the impression that for open road high speed tolling we are going to need better technology.

A: RF backscatter (passive) systems have been successfully deployed for many years in RF ID applications — straight identification, transmitting an identifying account number. That’s a perfect use of backscatter. Over the last several years these backscatter systems have also been deployed in plaza retrofit applications for more elaborate electronic toll collection procedures with reasonable success. True backscatter systems have an inherent advantage over active systems by having a more easily defined read zone which appears to make them an appropriate technology for use in a low speed, lane specific, plaza environments. However, most backscatter applications have still found it difficult to overcome cross-lane reads. Perhaps, it is because for safety reasons emitted power levels are limited to protect people. So given limits on the power of the system transmitter most deployments have found it necessary to provide batteries on the transponders to boost the return signal to achieve sufficient read accuracy levels even in these slow speed retrofit applications. These are therefore hybrid systems in which much of the inherent advantage of backscatter is lost.

Q: Are you saying if you need transponder power why not go active?

A: As far as open road high speed toll collection is concerned, longer communications ranges and higher data rates are needed to successfully complete the communications link. It is also necessary to locate the transponder as it passes through the toll gateway. The inherent physical limitations of RF backscatter systems retained in their hybrid counterparts limit their usefulness in open road applications. Some form of active RF communications is required to get both the communication range and data rates necessary to support open road ETC. Further, since vehicles are not channelized in an open road environment, tolling multiple lanes in a reliable manner requires some form of multi-tasking communications protocol — such as the public domain, time division multiple access (TDMA) protocol now being considered by ASTM as part of a national standard for ETC, CVO and a host of other ITS applications. The concept is easily understood by anyone who has ever used a CB radio. If you have your thumb on the mic button, you can only transmit, you can't receive! A TDMA protocol actually instructs each individual transponder in the read zone when to communicate with the reader. This enables it to manage collisions between signals and greatly enhances the read reliability in this application.

Q: There are claims of high accuracy rates with backscatter in the open road environment such as on 91-Express and the Oklahoma Turnpike. How should one interpret accuracy claims?

A: Yes. You have to be careful with the spin and there is a lot of it out there. For example when someone says ‘We did not have a misread in several thousand reads” one has to understand some of the nuances of RF communications to properly interpret that statement. Basically, if this technology reads the transponder, be it active or passive, it will almost always read it right. So it is a bit misleading to talk about just having a high percentage of accurate reads. The real challenge is to obtain reads of a very high percentage of transponders that pass through the toll gateway — and locate those transponders to specific vehicles so they can be tolled appropriately. This is clearly much easier in a stop and go, channelized plaza environment than in an open road, prevailing highway speed application. Further, there is too much focus on read accuracy. It is but one of many parameters that must be controlled to ensure vehicles are tolled properly.

Q: I hear toll authorities say they are requiring 99.995% read accuracy. Is that really obtainable and necessary ?

A: No. These are clearly requirements that have been imposed by some authorities in an attempt to obtain billing accuracy. However, billing accuracy is a system function and subject to several other monitors and processes. These limits should thus not be imposed blindly. For example what is clearly important is the ability of the system to reliably toll most vehicles that pass through the toll gateway. It is especially important to ensure that the system does not inappropriately toll many vehicles. A better systems performance monitor would thus be an upper limit on the number of vehicles that are inappropriately tolled.

Q: A “mis-toll rate”?

A: You can choose your terms. Transponder read accuracy is often secondary to the result, since the biggest difficulty is automatic vehicle classification because of complex toll schedules that were developed over the last several decades under manual toll collection methods. Further, for a performance criteria to be useful the authority must have some way of effectively monitoring performance, and sometimes this is not adequately considered.

Q: What is the bottom line on open road tolling? Are you saying it has to be an active system, that backscatter won't hack it?

A: Whenever we are talking about microwave frequency communications, safety is an issue — very low power levels are going to be mandated for health reasons. Using technology that is currently available and can be deployed in a cost-effective manner, it is clear that true backscatter RF systems do not meet minimum operating needs. In fact, at least some form of hybrid with battery boost of the backscatter signal is necessary to reliably support minimum communications needs for ETC applications in stop-&-go, retrofit applications at existing toll plazas. Prevailing speeds encountered in most open road toll applications further challenge the hybrid technology. Since hybrid applications retain none of the benefits passive systems, yet all of the challenges associated with their deployment, it appears to me that the smart choice for any open road application will be to deploy an active communication system. Also some form of multi-tasking protocol, such as the public domain TDMA protocol being considered by ASTM for the national standard, and the ability to dependably locate transponders as they pass through the toll gateway are also necessary.

Q: The Europeans and some US manufacturers think they can achieve adequate performance with their battery boosted backscatter systems?

A: There is a lot of specsmanship going around right now ! "Mine is better than yours." But I think they are stretching the technology close to, if not beyond its limits. I have worked with both and for the open road environment feel much more comfortable with the ATSMv6 active system. The Japanese did exhaustive tests of all the technologies and came to this conclusion too.

(Contact Daryl Fleming 770 734 9605 ISIFlem@aol.com. Comment on this viewpoint to TR Newsletter is welcome: tollroads@aol.com or fax 301 631 1248.)