2010/11/19

When will GNSS be chosen over RFID to Toll Roads?

In a recent email exchange over the various San Francisco proposals to deploy congestion pricing, a long-time road-tolling pundit remarked: "I'm not sold on GPS. It is expensive and unreliable as the Germans have found with their GPS-based Toll Collect for trucks on the autobahn system."

GPS (properly "GNSS") metering has come a long way since the 1990’s technology used in the 2003 Swiss and 2005 German systems. Companies from Germany, Austria, Italy, Canada and New Zealand all have reliable GPS metering in operation in the field and collecting revenue. Even the Toll Collect system resolved all their original reliability problems. In fact, Financial-grade GPS (FGPS) has been tested by Caltrans in San Francisco and has been reported here;
and this page also provides a URL to the original draft report from Caltrans.

Compared to RFID, GPS technology is more flexible, more extensible, and cheaper for large systems. If the only thing San Francisco will do in the next 20 years is one cordon, then it should use RFID by all means. But if this is the camel's nose in the tent (and you and I know it is), then GPS should be used. Had FGPS been used in London instead of cameras, the initial system would have cost 200M instead of 500M and the western extension 20M instead of 500M. RFID is perfect for one-time, non-extensible use. It is no longer appropriate for systems that will be extended or need to be flexible.


A recent paper given at the 2010 Slovenian Traffic Congress talks about 10 reasons GNSS is better than microwave
.  Actually there are 11, but the author left out interoperability (since the GPS signals are non-proprietary), and it was too late to change the paper before delivery.

E-ZPass has 22M transponders and 3700 toll lanes equipped with readers. I would suggest that the 5-year replacement cost for this system at  $1,000,000 per lane and $2 per vehicle (exclusive of operating costs) is $3.75B. Alternatively the 5-year replacement cost using GNSS road-use metering at $150 per vehicle is 3.3B.  We would have to study the operational costs, as well, of course, but let’s assume they are a wash (they are not, because E-ZPass is more costly.)

So already, for large systems, the cost of GNSS rivals that of RFID. The bill of materials for a FGPS-based device in volume is already approaching $100, and will go lower. RFID prices no longer decline: materials will go up, labor will go up, power will go up, construction costs will go up, maintenance costs will go up, right-of-way costs will go up.
But, GNSS prices will decline.  As the pressure for tolling more and more of our network increases, the case for RFID which is already founded more on habit than understanding, continues to erode. Only its installed base and fear of change preserves its legacy. Its economics is failing.

We will certainly make the switch from RFID to GNSS.  The question, now is only: "When?"
 
I think there is further evidence in the fact that Kapsch paid a measly $3.18 (about 1 drive’s-worth!) for each E-ZPass user or a paltry $2.9M for each of the 24 Toll Operators that are part of the E-ZPass group
. They bought the (captive!) customer base for a song. As richer telematics platforms for the connected vehicle provide more and more features, and as FGPS-based parking, insurance and road tolling become simple apps on these platforms, dedicated transponders and the hideous clutter of gantries will no longer be the gold standard for tolling.

But congratulations to Kapsch for picking the pocket of America as she sleeps.

2 comments:

mcapper said...

Bern

To give a 5-year cost differential is meaningless when the majority of the 3,700 E-ZPass lanes have been in service with original equipment for significantly longer than that. In some cases since the mid 90"s.

Martin

Bern Grush said...

Martin,

Thank you for your comment. My comparison was not a simple differential, since there is a huge sunk value in the installed base, as you correctly imply. My comparison was purely a replacement cost (I just added italics and bold for clarity), and as such is rhetorical for the current E-ZPass operators.

My argument would likely sway no one among the E-ZPass operators or any brownfield acquisition. Rather it is intended for those weighing RFID vs GNSS in a greenfield application such as contemplated by many states for VMT charging or congestion pricing in many locations. The specific trigger is the question: “Should San Francisco plan for a single one-shot deployment of a cordon? Or should it plan for a flexible, extensible solution in the face of rising congestion and failing gas-tax? How would you advise them?

If I am right and GNSS is applied appropriately, then in 10 or 15 years the E-ZPass operators will switch over via simple attrition since GNSS is super-operable (can interoperate ‘on top of’, using virtual gantries). The existing infrastructure will eventually be dismantled as was all full power, over-the-air analog television transmission signals in June 2009.

Bern