Making Better Lane Closure Decisions

In my last post we discussed the general importance of collecting work zone data before any temporary traffic control is installed. This was mentioned as an important lesson learned by several of the states that attended the peer-to-peer exchange May 21 and 22 in Iowa. They all wished they had accurate baseline measurements with which to compare.

In a later presentation W. D. Baldwin of HDR Engineering advocated the same thing but for different reasons. He said that by collecting baseline data before construction begins you will make better decisions about when to take a lane.

He suggested watching traffic counts. He uses PCE’s (passenger car equivalents). Most freeway lanes can handle about 1600 PCE’s per hour before queuing begins. If you measure counts over all lanes before you close one or more lanes, you will be able to see if the resulting counts will approach that number in the remaining open lane(s). If so, you should delay the work.

Many construction projects are bid with pre-planned work windows. These are traditionally based on historical data. But as we discussed in the last post, counts often differ significantly due to a variety of reasons. Mr. Baldwin explained that by measuring counts just before you start, you will make far better decisions about when to close lanes and when to open them back up.

This may be especially important for maintenance work and other short-duration work zones. Maintenance crews could place a sensor near the start of their proposed taper when they arrive at the work location. They could then hold their tailgate meeting with the crew to discuss that days’ work while the Traffic Management Center monitors counts. As soon as counts are low enough, the TMC could call the supervisor and tell him it’s OK to close the lane.

The other side of this same coin is some form of variable work windows. Utah DOT is considering a system of incentives or disincentives where the contractor decides when to work and when to stop. They are penalized when speeds drop below a certain level, or when queues exceed some agreed length, or some other measure is exceeded.  Both the DOT and the contractor see the data in real time so everyone is on the same page. It gives the contractor the opportunity to get more done when traffic thins out.

In cities like Los Angeles and San Francisco work windows can be as little as 4 or 5 hours. Traffic control takes a good hour to set up and tear down leaving the contractor as little as 2 hours to work. In cases like these, if a variable system got him just one more hour twice a week, it would equal a 20% increase! That would result in faster completion and reduced exposure – increased safety and reduced costs – a “win-win” in anyone’s book!

The Need for Baseline Data

This past week I spoke at an FHWA Peer-To-Peer Exchange on Work Zone ITS. I talked about work zone data: performance measures during and after a work zone is installed, how to collect the data, and lessons learned from past deployments. The DOT folks in attendance were very involved in the discussion and made several rather astute observations. These fell into three groups or themes which I would like to share with you in my next few posts.

The first theme revolved around the need to deploy work zone ITS prior to any other work being done. Work zone ITS systems are used for a variety of reasons but nearly all of them benefit from good baseline data. Most of the meeting participants had used historical data and said it was better than nothing. But that current data would have been more useful.

Historical data are averages. They look at AADT over several months. Those may have been the same months during which construction will take place, but chances are they were not. Furthermore, changes affecting counts are taking place every day. Special events, work on other routes that change traffic flows on your route, weather, economic conditions, and many other changes affect the overall and hour-by-hour counts passing through the work area.

How do you know if the work zone is causing traffic delays if you don’t know what it would be without the work zone? If your system is being used to measure and display delay times, what will you base that on? Delay times are best used on commuter routes. Commuters know better than we often do how long it should take them to get from point A to point B. If we want to be accurate about the delay times we display, we must have an accurate baseline.

There is also the possibility that the algorithm used to calculate delay times results in numbers that are always a little higher or lower than the actual delay times. It is far easier to prove that before a work zone is installed than once the work begins.

If you deployed the system to reduce the number of crashes or incidents, you will also need a baseline. The simplest way of setting that is to deploy your system early and measure the number of times speeds drop below a threshold, say 30 MPH. Then compare that to the numbers once the work zone is in place.

To evaluate a system’s effectiveness, regardless of your deployment goals, you need baseline data. And you don’t have good baseline data if it wasn’t collected just prior to the project start. Everyone who spoke up at the meeting last week had experiences where they wished they had the system in place earlier to collect baseline data.

So as you plan your next project where work zone ITS will be used to mitigate traffic impacts, be sure to bring the system in a few weeks early. It gives the contractor time to iron out any issues and, more important, it will provide you with the very latest counts, speeds, volumes, or whatever you choose to measure. By comparing current baseline data with the data after the work zone is in place, you will make better decisions about work zone design, work windows, and much more.

Developments in Work Zone Video

OK, it is time for a mea culpa. That’s right, write this down because I am here to apologize. In my August 27 blog entitled “Cameras or Sensors?” I took the side of sensors. I argued that sensors are much less expensive than cameras and provided far more information. To my mind that has not really changed but video has come a long way in recent years and deserves a second look.

Recently, one of the most important developments for our industry is in quality digital video recorders or DVRs. They have become relatively inexpensive yet offer an incredible amount of storage. Before these became available, you had to react to alarms (provided by sensors) when the problem occurred to view the video. You wouldn’t necessarily understand how the problem developed but at least you could see what the problem was and formulate a response. So, at that time, the value of video was more reactive.

Now you can go back in time and watch how the problem developed. What happened to cause it? Was it an errant driver? Or did traffic control contribute to the problem? In other words, you can use stored video to improve upon your traffic control. It has become proactive, rather than simply reactive.

Wireless service has become more dependable and 4G service offers good bandwidth if you need streaming video but it can be expensive. And in most cases you won’t need that. A few frames a second will provide all the real time information you need to understand what is happening in a work zone. Use the money you save to add more devices or to use these systems in additional work zones.

Another change is the variety of cameras. Most manufacturers offer good 12 volt cameras now. And you may choose from color, black & white, infrared, or some combination of all three. They run from very basic “grainy” video to high quality megapixel video. The more basic the video, the more you can store on the DVR. High quality video is great, but takes far more storage space. You must evaluate that trade-off between video quality and storage. Most DVRs will begin recording over the oldest video when it is full. So the question becomes, “When an event occurs, how soon must I download the video before it will be lost?” In most cases you can design a system with 30 days storage, giving you plenty of time to find out about a problem after the fact and still go back and view it.

Cameras now offer a few more bells and whistles that may also be of interest. They can be set to scan a work area continuously. Or they can be set to focus on a particular area during morning drive time and another area late at night. This feature might improve security where theft or vandalism is a problem. Alarms or flashing lights can even be triggered when movement is detected late at night.

This feature could also be used to leverage your work zone ITS dollars. If traffic is heavy in only one direction at a time, one camera could be used to view both directions of traffic. In the morning it could view the traffic heading into town and in the afternoon it could be set to switch automatically to the outbound direction.

Cameras can still be manipulated manually. Any time a new alignment is installed, particularly lane shifts or crossovers, you should monitor traffic through it to be sure drivers understand what is now expected of them. If they handle the change well, you can pat yourself on the back for a well designed work zone. But if problems develop, you will know you need to make adjustments.

Cameras are still several times more expensive than sensors. Sensors must provide the data backbone of any good work zone ITS system. But stored video can be another useful tool in your tool box…especially when analyzing problems and making improvements in traffic control design.

The Value of Real-Time Data

Today I would like to discuss data. Everyone agrees it is important, especially in light of the focus on performance in the new highway bill. If we don’t collect data, we have no way of evaluating performance. But what data should we collect? And how can that data be used to both evaluate performance and help to improve that performance over time? We need some data now to make decisions in real time. But is that the same data we need when the project is complete and we are looking at ways we could improve next time?

This is a complex subject. Over the next few blogs we will consider three issues related to data:

1)      Data is a key product of work zone ITS. Always require access to it, in real time.

2)      Speeds over segment versus speeds at key points in work zone.

3)      What are best metrics for wz performance?

Today, we begin with data and the importance of access. Work zone ITS systems generate a large amount of data. At the top of the list is traffic speeds, counts, and volumes. There are the logs of messages displayed on each message sign and the alerts sent as emails or texts to key personnel. There are the logs showing who accessed the system and who chose to override the message sets. There are also the logs showing voltage levels and other measures of device status. This list boils down to:

1)      Work zone performance measures.

2)      System outputs.

3)      Security and system overrides.

4)      Maintenance.

When choosing to include work zone ITS in a project to mitigate traffic impacts, our goals vary but they generally fall into two categories. Our first goal is usually to maximize throughput and safety. Any other goal is going to be in support of this first goal. They will be to measure the effectiveness of the system, to collect data for use in designing future projects, etc.

Data needs are either short term or they are long term. The long term is used to look back once the project is complete to evaluate the performance of the system, evaluate the other work zone traffic control or staging, or other uses that are non time sensitive. There are also a wide variety of short term needs.

The most important of these is “feel”. By digging into the data you get a feel for the location, time of day and duration of traffic “events”.  That makes it far easier to make effective adjustments in staging or traffic control.

Feel also helps you to eliminate false alarms and/or verify all alarms were legitimate. Once you check several low speed alerts, you will know they are accurate. If they are not you will know which sensor locations should be adjusted to avoid false alarms from slow moving construction equipment or other anomalies.

Raw data helps you adjust alarm parameters until only the true emergencies are treated as such. Agencies always set the trip speeds high to begin with, then slowly reduce them until they are comfortable.

Access to real time data allows agency and contractor personnel to get their hands dirty. They can drill down into the data and get a better feel of how the system works. This will result in more effective use of the system on the current project and on future projects.

This is only possible when you have real time access to the data driving the system. So be sure to specify it the next time you include work zone ITS in a project.

Work Zone Performance Measures

I just read an FHWA study entitled, “A Primer on Work Zone Safety and Mobility Performance Measurement” written by Jerry Ullman and Tim Lomax of TTI and Tracy Scriba of FHWA (FHWA Report # FHWA-HOP-11-033). Since SAFETEA-LU we have been wrestling with how best to measure and evaluate work zone performance. Much has already been written, but this report takes the practice to a new level. In less than 50 pages it explains the need for measurement, the types of data that might be used and reasons a state or local agency might choose one or more of them.

I won’t try to create a Readers Digest version of the report here. You will have to read it yourselves and I strongly recommend that you do. But I would like to share a few random thoughts that occurred to me as I read it.


  • They contrasted project level measures with agency/program level measures. Most would agree that work zone ITS systems can help with both. But our focus will always be on the acquisition of project level data.
  • The measures chosen will vary widely from agency to agency for all of the reasons listed in the report. But also because agency goals vary widely. Agencies with urban, congested roads will be more concerned with volumes. Agencies in more rural areas with lower capacity will be focused more often on delays. Also, no two agencies have the same categories or detail of data available to them. Performance measures must be tailored to best reflect the goals of the agency while measuring what needs to be measured.
  • Beginning on page 41 they discuss methods for estimating queues and delays from spot sensor data. What do you system manufactures think about this? Does it agree with your experience? Or would you change something? Please tell us what you think.

Finally, on page 30, they offer some very good advice when using spot speed sensor data:

  • Ensure that sensors will exist within the work zone and upstream for a distance greater than the anticipated length of congestion and queues that may develop.
  • Deploy sensors as closely as is practical and affordable, to increase travel time and delay measurement accuracy.
  • Ensure that traffic sensor spot speed data will be archived for use in work zone performance measure computations.

This fits with my experience. They are three important lessons learned. We often wish we had more data, but I have never been involved with a project where we felt we had too much data. (See “Cameras Versus Sensors, August 27, 2012)

Please read this valuable study and let us know what you think. And, please, pass it on to your fellow practitioners.

Cameras or Sensors?

I don’t claim to be an expert on this subject. In fact, I hope to learn more about it from you. But I believe that when you are budgeting your work zone ITS, you are almost always better off putting most of the money into sensors rather than cameras. You only need so many cameras and you probably don’t need any at all. But I’m not sure you can have too many sensors.

In my experience, cameras cost at least three times more than sensors. The device itself is more expensive because it needs a larger trailer, more solar and batteries to power it, and a taller tower. It also costs more for wireless service because it requires much more bandwidth and it requires it continuously, not just for short packets of data. In some cases cameras cost more than 5 times as much as sensors. So, for your money, you can get three to five more sensors for every camera you give up.

We need to better understand both tools before we can discuss the budget. When do we need cameras? Most of the time cameras are used to verify the data from the sensors. Is traffic really stopped? Or was this a false reading? But if you have enough sensors, the ones on either side of the one you are verifying will quickly do it for you. And, they will do that automatically. Just because we can check the data obsessively doesn’t make it a good idea.

But this is not to say we never need cameras. There are some good reasons to have them:

  • Publishing video to a webpage helps drivers check conditions quickly and raises public awareness about the work zone.
  • Emergency response may be improved by viewing the incident first to know just how bad it really is and to direct responders to it in the most efficient way.
  • Some DOTs may want a video record of the work zone traffic control. This is still controversial but whether you agree with the idea or not, it is still a legitimate reason for cameras.

A reason that is less legitimate is the use of cameras to look at ongoing work or to look at locations on the jobsite without having to drive there. This does save time and money, but not enough in itself to justify the cost.

These are some of the questions you should ask yourself when evaluating the need for cameras:

  • Are there permanent cameras at or near the project that can be used?
  • Is the job several miles long? The longer the project, the more important video becomes to groups like emergency response.
  • Will there be several stages to the work? As construction progresses and traffic control is changed, it is important for drivers to know about those changes in advance.
  • Is public outreach a part of your TMP strategy? If so, a web page with video might be an important tool.
  • Will there be lane shifts, narrow lanes, or a significant reduction in capacity during construction? Any of these conditions may result in an increase in collisions. The more problems you expect, the easier it is to justify the expense of cameras.
  • Is this what FHWA calls a “major” project? The larger and more complicated the job, the more likely you are to benefit from CCTV.

Cameras have their uses. But next time you are planning a job, please ask yourself these questions and consider your goals for work zone ITS before you decide to add them to the project. You may be able to do the same thing more efficiently by just using sensors. What do you think? Have I left out important questions? Is there a side to this I have not considered? Make your comments now and keep the discussion going!