HOV Lanes: A New Study Finds They Work Better Than Motorists Think
An HOV lane with a buffer separating carpoolers from regular traffic.
High Occupancy Vehicle (HOV) lanes, also known as carpool lanes, have gotten a bad rap according to new research by a team led by ITS professors Michael Cassidy and Carlos Daganzo.
Their paper, “Empirical Reassessment of Traffic Operations; Freeway Bottlenecks and the Case for HOV Lanes," disputes earlier findings that examined six freeway sites in the Bay Area and concluded that HOV lanes were responsible for increasing congestion and should be eliminated.
“That (earlier) study was very compelling to us,” explained Cassidy, Professor of Civil and Environmental Engineering. “If HOV lanes were causing the (congestion) problem, maybe there is a silver bullet: if we got rid of the HOV lanes perhaps we could mitigate a good deal of congestion.”
But when the researchers, which also included Koohong Chung of Caltrans and Kitae Jang, a researcher at ITS, looked at the earlier data and collected more information, including videotapes of the areas to examine how traffic speeds and flows changed over time, they were surprised to find just the opposite effect.
“To the contrary, and quite remarkably, the HOV lane seemed to increase the capacity of the bottleneck that was videotaped--even though that lane was underutilized,” they wrote. In other words, the HOV lane was helping traffic in all lanes to move more quickly—not more slowly--through the bottleneck.
Misplaced blame
What the researchers found at the sites they examined was that the HOV lanes were not causing delays or queues. Instead, in virtually all cases, traffic backups, or queues formed first at bottlenecks, and almost always for reasons unrelated to the HOV lanes, such as an accident, construction, or increased traffic during the rush hour.
Even at a site on northbound 880 in Alameda where Cassidy and Daganzo were pretty certain that the previous researchers had got it right, a closer evaluation exonerated the HOV lane. Examination of that site over several days and with the benefit of videotape found that slowing travel speeds coincided closely with the activation of the HOV lane at 3 p.m. On closer inspection, however, they discovered that the bottlenecks and congestion formed with the increase of rush-hour traffic, not the activation of the HOV lane.
“In this case, it’s like Spiderman showing up at the scene of the crime so quickly that he gets blamed for the crime,” explained Cassidy. “Here the HOV lane is going into service right about when you’d want it to—when the bottleneck and congestion forms—and then gets blamed for the congestion. On this particular location the activation of the HOV lane is just dead perfect. So here Caltrans should be getting kudos.”
Additional solutions
The authors are quick to point out that their research does not mean that HOV lanes never cause congestion. An extremely underutilized HOV lane, one that is virtually empty, would be a disaster, explained Cassidy. But in some—perhaps many—cases drivers in ordinary, or low occupancy vehicles (LOV) lanes, benefit from HOV lanes, even as they curse them.
“An HOV lane, even one that’s somewhat underutilized, can increase bottleneck capacity because it diminishes disruptive lane change maneuvers,” explained Cassidy. “If you don’t have as many people moving in and out of the median lane, that can be advantageous. We call that the ‘smoothing effect’ because it smoothes the bottleneck flow.”
Although the authors found no convincing evidence that the HOV lanes they scrutinized in their research caused backups, HOV lanes can cause congestion at other sites. But even at some of these sites, modest alterations could relieve the problems.
Caltrans should probably change its approach to HOV lanes. “Currently there is a sort of one-size-fits-all approach,” said Cassidy. A closer look at each freeway’s bottlenecks might reveal that the hours of HOV operation could be tweaked to work more effectively.
Another modest alteration involves allowing low-occupancy vehicles to periodically move into the HOV lane near a bottleneck. “You wouldn’t let them in continually, only periodically, so as to maintain a queue of short length of a few hundred or a thousand feet in that HOV lane.” Changeable message signs could be strategically placed to control when and how frequently to let low-occupancy vehicles into the HOV lanes.
This would not penalize the HOVs very much, Cassidy said. Although drivers in the HOV lanes would have to slow down when they hit the queue, it is only a short queue, he explained.
“The advantage might be twofold: you might get this reservoir of demand trying to pump its way through the bottleneck, and you might get the smoothing effect that we saw in the video data.”
The researchers also point out that an underutilized HOV lane will cause queues in the adjacent low-occupancy lanes to grow faster and longer over the freeway. This effect can create congestion problems if the added queue blocks busy ramps. The researchers, however, have formulated theories showing that on freeways where queues have room to expand, these negative effects tend not to be significant. For additional discussion, please see the Cassidy and Daganzo's second report, “Deploying Lanes for High Occupancy Vehicles in Urban Areas”.
Cassidy and Daganzo's research on HOV lanes is the topic at the April 20 ITS Friday Transportation Seminar.