BAY AREA:Rough going for Bay Area ET

In the early hours of Dec 20 the south Bay area of San Francisco clattered to the thucka-thucka beat of TV news helicopters getting their stories of “FasTrak gridlock” for the evening news, while the radiowaves carried live reports of the traffic tie-ups, with the wire-services not far behind. On the ground reporters were out there in the stalled traffic getting the reactions of despair and annoyance of motorists stuck there, unable to get to work.

Caltrans had run into the classic timing problem on the Dumbarton Bridge by attempting to introduce electronic tolling (ET) before enough motorists had transponders. Such problems are particularly acute at a small toll plaza – and this one had just six toll lanes.

That weekday morning Caltrans took away one cash lane for an ET-only lane, and all hell broke loose. Traffic backed up 21km (12.5 miles) and a morning commute trip that normally took about 30mins ended up taking an hour and a half. Caltrans was beset by angry motorists and deluged by negative news reports.

The problem – far too few motorists with transponders for a conversion of a toll lane to ET-only on a small, already tightly stretched toll plaza. The Dumbarton bridge is the furthest south of the eight toll bridges in the Bay area. A reasonably modern 6-lane concrete structure it links I-880 on the east side of the Bay in the Fremont/Hayward area with US-101 (Bayshore Fwy) around Stanford, so it is a major commuter route for Silicon Valley employees.

In an area thicker with PhDs and other brainy and enterprising people than perhaps anyplace else on earth, the simplest arithmetic, and a rather common managerial judgment call on trade-offs apparently eluded Caltrans on this one.

Arithmetic

With six toll lanes a decision to convert one to ET will clearly slow cash paying traffic unless the numbers of cash-payers have been reduced to five-sixths of the total, meaning that at least one-sixth of motorists need to have ET-transponders to maintain cash toll lane throughput unchanged. You go out before the ET-switchover and eyeball the vehicles, look for those ET tags on the windshields and count. That’s the drill at most toll agencies as they go to ET. If less than a sixth of vehicles have transponders you know there is going to be trouble in the cash lanes. The more the number is below the proportion of cash paying toll capacity being eliminated the more the potential trouble on switchover.

There are also some beautiful simulation applications for your PC (for example from Long Island based KLD Associates, Vollmer and TransCore) which will help quantify the results of different scenarios. They can be rerun with different plaza configuration and throughput assumptions, and have random traffic variability fed in. They will present an animated moving picture of what is likely to happen in different toll plaza arrangements. From the summer on however Caltrans was under the political gun to get its system up at all plazas by year’s end no matter what, so there was no time for such sophistication.

At some point in a planned switchover with insufficient tags out there the management decision is: (1) do we want to take the heat from backups of cash-payers, warn motorists ‘Sorry be warned, you cash payers are going to have longer waits’ and use those backups to promote transponder uptake, or (2) let’s just go mixed-mode for the time being so all six lanes continue to take cash and a single lane handles ET as well as cash or (3) defer the ET-startup until marketing has gotten enough transponders out there on windshields, or (4) you lay some asphalt for an extra toll lane so no take-away is needed.

None are easy. Each has its downsides: (1) Brings down a lot of motorist wrath, negative news coverage representing you as bunglers, incompetent etc (2) mixed mode denies the motorists with transponders a lot of the quick ride that they’ve been promised because they are stuck in the regular traffic like everyone else – indeed they are somewhat worse off because at least the cashpayers can choose any lane (3) you’ve got a chicken and egg problem of how to persuade people to get tags if you aren’t reading them, and a credibility problem of when you REALLY are going to start ET (4) building an extra lane will be costly, there may be space issues, and charges of wastefulness for building something needed for perhaps only a few months.

The smaller the toll plaza and the tighter it is already stretched the more difficult these dilemmas. [Note the old toll managers rule of thumb of three cash toll lanes needed per travel lane would have dictated 9 toll lanes at the Dumbarton with its 3-travel lanes in the toll direction, not the 6 toll lanes it has!]

A Caltrans officer told us that the basic mistake they made at the Dumbarton was to attempt to estimate potential ET usage by a zipcode analysis of ET accounts in the east bay area. They found that about 10k people in Alameda county (Livermore, Pleasanton, Fremont, Hayward etc) had tags and therefore thought there would be plenty of customers with tags coming over the Dumbarton.

Mistake on tag numbers

It turned out that by 8am Dec 20 only 135veh/hour were using the ET-only lane, which of course has a capacity of 1,000 or more/hr. With traffic backing up I-880 all the way to CA-92 in Hayward there was a major crisis. The problems on the Dumbarton threatened to block access to the San Mateo bridge 14km (8mi) up-Bay – the other crossing route from the east Bay to Silicon Valley.

“Those people we thought would bring transponders to the Dumbarton, it turns out they must go north and use their transponders on the Bay Bridge, or one of the other bridges further north. They weren’t people who use the Dumbarton. That’s how we screwed up.”

In the middle of that first morning rush hour around 8am Caltrans threw in the towel and put a toll collector back into the booth covering the ET-only lane and changed it to mixed mode. It will operate mixed mode now until Caltrans is quite sure there are enough transponder equipped vehicles to justify the dedicated lane.

Caltrans says that they have had complaints from transponder users that they have lost “their lane” but the experience of Dec 20 has convinced them there must be mixed mode until the tag population on the bridge grows considerably. Indeed Caltrans is scrambling to install ET in the 5 other lanes on the Dumbarton so that all can be operated mixed mode.

Short of techs

As recently as last summer the Caltrans plan was to equip all approx 53 toll lanes on the seven bridges with ET to allow a flexible mix of mixed mode and dedicated ET operations and to fire up through the second half of 2001. But the smooth rollout of ET at the Golden Gate Bridge in July prompted a decision at the top of Caltrans to have ET operating at all bridges by the end of 2000, not 2001. Contractor Adesta has been unable to employ enough qualified technical workers – electricians and radio engineers – to wire up all the lanes this year.

So the decision was made to go with mainly dedicated lanes this year, since this was all it was possible for the contractor to do. At first it didn’t go too badly. A FasTrak-Only lane on the Benicia-Martinez bridge opened Oct 25. With 9 toll lanes total, taking one didn’t cause serious problems in the cash lanes, perhaps also because some of the patrons with ET tags for the parallel Carquinez were atracted to the ET-only lane on the Benicia.

There were signs of trouble however on the Richmond-San Rafael bridge Nov 15. That has only 7 toll lanes and for the first week there were big backups. But they didn’t go back to any major interchange and as the days passed ET-usage increased and the cash backups decreased. Here synergy with the Golden Gate bridge and its large base of transponder-equipped customers helped alleviate the early problems. A dedicated ET lane plus a mixed mode lane opened on the biggie, the 240k veh/day San Francisco-Oakland Bay bridge Nov 29. Slight increase in normal aggravation, they said but the difference wasn’t huge and within days things were back to normal and perhaps even a tad better. Success!

Caltrans had two more bridges to fire up as we go to press – the San Mateo and the Antioch Dec 27 – and both these will open mixed-mode. Traditionally the argument is: if you have no dedicated ET to give tag users a break, then you have to take ET in all lanes, or you are leaving the ET-payers worse off than the cash payers with a lesser lane choice. But if your contractor doesn’t have the technicians then you’re stuck with what you have.

50k tags

Caltrans has 50k tags out there as of Dec 20. ET-users get a 15c discount (7.5%) on the $2.00 standard car toll, though that is presently due to be discontinued Dec 2001.

The Caltrans implementation of ET has been the most troubled of any to date. MFS (now Adesta) was contracted to develop, install and operate an ET system on the Bay area bridges as far back as 1993. In 1996, though behind schedule it was said to have narrowly failed tests with the expectation of full operation in less than a year. It has had an ET lane operating on the Carquinez bridge since Aug 1997 but has been unable to get a system working satisfactorily, so for three years extension of ET to the other toll bridges in the Bay area was deferred.

Caltrans project manager Charles Price told us that the system as deployed is now working well. The project has cost about $50m, Caltrans having approved additions to the original price of $35.9m in 1999. A senior officer said in the fall that the department overall had “mismanaged” the program. In addition the contractor MFS/Adesta’s reputation has been considerably sullied, and its bank accounts depleted, by the job.

Twin antennas

An unusual feature of the RF data links in the Caltrans installation is a twin antenna arrangement in each of the ET lanes. A single antenna is normally deployed which both sends a signal and receives the response from the customer’s transponder. Price says that at the Caltrans installations they were unable to get a single ‘transceiver’ (send and receive) antenna unit to work adequately with transponders on a variety of different vehicles.

“We got cross-reads (from adjacent lanes), multiple reads, all kinds of radio signal problems. We couldn’t get adequate performance the way it was first set up (with a single transceiver antenna.) With separate antennas we can tune the coverage much better.”

The antennas are both in the toll plaza canopy, the transmit one just ahead of the middle, the second or receive antenna close to the exit end of the canopy.

Indeed little remains of the ET system originally designed and installed by MFS in the Carquinez bridge in 1996. That involved, for example, laser profilers as the principal tool for vehicle classification. Now installed for this purpose are quite conventional axle counting treadles. No profilers. At one time MFS got much positive publicity for a special treadle design it devised with a subcontractor at its own expense for Caltrans. It used a diagonally laid treadle to distinguish single from dual tire vehicles. Not there.

What remains unchanged is a basic electronic toll system including readers and transponders supplied by Sirit of Toronto, the company which took over the Texas Instruments business that designed the basic passive backscatter system that is known as a Title 21 standard, and is the principal ET supplier in California.

All ET lanes are signed FasTrak, a trademark of southern California’s TCA toll roads but used on all toll facilities in the state to designate electronic toll lanes. Five (soon six) separate entities collect tolls in the state and use ET. But as far as motorists are concerned there is a single system, FasTrak, because the five agencies all honor other agencies’ transponders and remit monies collected to agencies holding the account.

The ET-only toll lanes at Caltrans facilities are signed for 25mph.They say some have been tolling over 1500 veh/lane/hour on the busy Bay bridge.

Caltrans Bay area bridges collect some 321k tolls per day for nearly 640k daily trips (tolling being westbound only), making it the third largest crossing system by toll volume in the US – just behind PANYNJ’s 339k and NYC’s MTA B&T’s 792k.

Benefits

Motorists will have more trouble seeing benefits from ET at the Bay bridges than at some other toll facilities. Congestion relief will be limited. That is because many of the congestion problems are in the travel lanes themselves, not in the toll lanes at the plazas. The Oakland Bay bridge for example carries close to 25k veh/day/travel-lane (250k/10) compared to about 22k (300k/14) at the George Washington Bridge NY/NJ.

Signal metering is widely used on Bay area bridges to control the entry of vehicles onto bridges, and ET users get held up by the signal meters just like everyone else, detracting from the experience of getting a quick ride-by the congestion which is found on most east coast toll facilities where ET-mode share is now generally over 50%.

Caltrans has a policy of favoring HOV travel with toll plaza bypass lanes at many of the toll bridges. This further complicates giving preference to transponder users. On most toll facilities transponder users can be made ‘top dog’ whereas in the Bay area this status seems likely to continue to be held by carpoolers.

Finally in the moderate balmy climate of California, it’s a lot less of a drag winding down the window – it may already be open! – to hang out a pair of dollar bills than in New York where there may be sub-freezing winds blowing, or alternatively 90 degree/90 percent humidity outside. (Contacts Charles Price ETC Program Manager 510 286 4478, Greg Bayol PR Caltrans 510 286 6169)