In Oregon, the gas tax funds most road construction and maintenance. Seven years ago, the legislature realized that the state faced the prospect of a sharp drop in gas tax revenues due to improving vehicle fuel efficiency and the introduction of hybrids, as well as the political impossibility of raising the gas tax per gallon. So it passed a law creating a task force to develop a new road revenue system alternative to the gas tax and directing the state’s department of transportation to run a pilot project. After investigating 28 different potential revenue sources, the task force decided to explore the possibility of a mileage fee, such that drivers would pay a “fee for service” rather than a general tax unrelated to their actual road use. The system had to be affordable to implement statewide, enforceable, provide redundancy in the event of technology glitches, and be easy on motorists.
As James Whitty, the state’s manager for the pilot program, wrote in a report released in November 2007:
While principally designed to replace the gas tax over time, the Oregon Mileage Fee Concept also provides an electronic platform for creative applications of congestion pricing to manage levels of traffic during peak periods of driving. In other words, the concept accommodates creation of multiple “zones” that allow not only local option but also various pricing methodologies.
I discussed the technical challenges with engineering professors David Kim and David Porter, of Oregon State University, who led the development of the pilot project.
Constraints
The developers were working under several constraints, including:
- Since the field test was of limited duration, the on-vehicle devices could not be manufactured into the vehicles.
- They were not permitted to make any permanent changes to the volunteer motorists’ vehicles.
- They were not permitted to alter the software of existing fueling stations’ point-of-sale (POS) systems.
- The system at the service station could not involve any extra human interaction or effort.
Therefore, the field test retrofitted temporary, prototype on-vehicle devices into vehicles privately owned by volunteer participants, employed lab-generated data transmission technology, and jury-rigged modifications to existing fueling station systems. The on-vehicle devices used GPS receivers to determine in which zone each mile was driven.
According to Kim, the fourth constraint was “the hardest thing we had to deal with” because service stations have multiple pumps in close proximity and not every vehicle was equipped with one of these devices. That made it challenging to ensure that each pump would read the data from the vehicle closest to it and not from one at a different pump. “We referred to that as vehicle-dispenser association,” he says.
To make the identification more reliable, they installed a main reader at the gas station plus an additional device on each pump. When a vehicle equipped with a device pulled up to a pump, the main reader would instruct all the pumps to broadcast a signal that was then captured by the device in the vehicle. “By looking at the strength of that signal,” Kim explains, “we made a probabilistic determination, using an algorithm, and that helped us better associate a vehicle to a dispenser.” Because they were not allowed to make any permanent modifications to the equipment in the service station or to the volunteer vehicles, they discarded suggestions to install readers in the nozzles or in the ground.
The pilot project had only three zones — out of state, in Oregon but outside the Portland metropolitan area, and inside the Portland metropolitan area — which were managed by the software in the device in the vehicles, though the device was able to handle more zones. If the system went live, it could potentially involve many zones —whether for congestion pricing or other purposes — which might be changed and updated frequently. That would require an interface with the in-vehicle device, which could also be used to upload software updates. One way to do that would be to connect the devices in the service stations to a network, so that they could receive the updates and then transmit them to the devices in the vehicles.
Tampering and Fraud
According to Whitty’s report, currently motorists would have little incentive to tamper with the devices because it would only save them a small amount, but this could change if the state were to eliminate the gas tax and switch completely to a mileage fee. Kim and Porter’s tests did not explicitly consider the possibility of tampering. However, they designed the system to have a central database of odometer readings updated at each transaction. This allows it to flag vehicles that are not accumulating any mileage. That could only be due to one of three reasons: the vehicle is not being driven, it is being refueled out of state, or someone is tampering with the device.
However, to protect drivers’ privacy and keep it simple, the system does not store the GPS receiver’s track log. This can make it challenging to prove tampering or fraud. If the state were to implement this system, it would have to balance privacy concerns with enforcement requirements. Some countries that have implemented similar systems for congestion pricing, such as England and the Netherlands, have set time limits on data storage. “You could figure out a time window during which this data would be kept so that you could audit the system or to identify possible tampering or malfunctions,” says Porter. The time interval would depend on how frequently people fuel up, on average, in a given area. Disputes of charges, Kim points out, would be much harder to resolve and would require more data. “If a device were built into the car, we could always match the total mileage with the odometer and they would have to be relatively close.”
Interface with the Vehicle
Kim and Porter had problems trying to interface with the diagnostic system in some vehicle models and were not able to cover the whole spectrum of standards. Looking ahead, they say, instead of retrofitting devices into vehicles, it would be best to work closely with car manufacturers to take advantage of the technology that they are already building into them — such as navigation systems, bluetooth, Wifi, and RFID — to calculate the mileage driven and to collect and transfer the data.
In the pilot program, the majority of the devices obtained the mileage from the vehicle, through the diagnostic data port, and used the GPS receiver to determine where that mileage was accumulated, while some devices used the GPS receiver to perform both of these functions. The pilot program did not use cellular technology, mostly due to privacy concerns. Connecting the in-vehicle devices to cell phones would greatly facilitate data collection and software/map updates; however, there would be unacceptable gaps in geographic coverage.
Division of Labor
What was the division of labor in developing the pilot program? What was developed specifically for this project and what was commercial off-the-shelf (COTS) technology? “We did not have any COTS,” says Porter. “That was our first realization. Initially the state was under the impression that these devices were already available and that they could just slap them in the cars and were going to have exactly what they needed. Everything had to be developed, especially from the cars’ perspective, from scratch; we developed the requirements based on the boundaries that the state had set. We hired a developer and we had a company put it together. On our side, the other component was the central database, which we developed here at OSU and kept here for the duration of the pilot. We received and stored all the transactions from the gas stations and generated a bunch of reports for the state to do what they call the “true-up” — to reimburse the stations or collect money from them, depending on the difference between the gas tax that they collected and the user fee transaction. The components that went into the device were COTS, but the device itself was a custom device that we developed for the purpose of the project. The software that coordinated the transactions between the POS systems, the odometer readings, and the central database was developed by OSU graduate students.”
Scaling Up
What would it take to scale up the system from the pilot to state-wide adoption? According to Kim, the design of a multi-state system has to be considered. According to Porter, the key would be developing the relationship with car manufacturers, as well as enabling vehicles to communicate with the roadside infrastructure. “Why re-invent the wheel when we could take advantage of that?” he says.
Kim cautions against trying to achieve too much too quickly, especially with regards to the geographic zones. “We don’t know exactly when people drove in a given area, so there can be many disputes. Our system had it by day, rush hour, peak and off-peak. Those are pretty broad time periods. If the rates or the zones changed, the device would not be able to tell how many miles were driven under the old ones and how many under the new ones. It could get really messy! If this thing is going to be implemented in the next few years they have to keep it simple. If they try to put too many bells and whistles — and deal with such things as congestion pricing, changing zones, or even collecting traffic and travel times, as some have suggested — that is going to run into too many problems.” Right now, he points out, the gas tax does not achieve any of these other policy goals, so even a simple system would be a great improvement. (Additionally, the more numerous and smaller the zones, the more often drivers would cross zone boundaries, and the more critical would be the requirement for continuous GPS reception — which is lost in tunnels and in-door garages.)
The only way to handle complex zones and rates would be through a central system that would continuously allocate the miles by zone and rate. This would require either storing the track data and downloading it at the gas station — which creates problems for privacy and, possibly, bandwidth — or real-time data transfers, which would require continuous network connectivity.
Questions
This Oregon experiment raises many interesting policy and technology questions. Here are a few:
- Could such a system take full advantage of GPS technology while still protecting privacy?
- If this system were to become mandatory and universal, would it greatly expand the market for location-based services (LBS), similarly to the way that GPS receivers in phones did?
- Would vehicle owners be required to buy and install the equipment, at their expense?
- What would happen if a device malfunctioned or froze up?
- Who would maintain and update the fee zones?
By: Matteo Luccio, President, Pale Blue Dot Research, Writing, and Editing, LLC
