Variable Speed Limit Systems – Revisited

In a post on April 24th we discussed a recent webinar on variable speed limit signs. The speakers voiced their disappointment in the technology and found very little if any benefit from their use. But now a new study by the University of Missouri saw far better results for VSL signs used in work zones.

“Evaluation of Variable Advisory Speed Limits in Congested Work Zones” by Praveen Edara, Carlos Sun and Yi Hou found far better, but still mixed results when using Variable Advisory Speed Limit signs in work zones.

As we have known for some time, the VSL results depend to a great extent on the algorithm used. In the Missouri study the original algorithm used in the field resulted in shorter queue lengths and reduced speed differentials. In fact the maximum speed differential was reduced by a remarkable 10 MPH. However it also reduced throughput by 7 to 11% and travel time increased by 4 to 8%. So the results were mixed.

They then experimented through simulation. Using the same traffic data they tried two other algorithms. One smoothed over a one minute period and the other over a 5 minute period. The original field algorithm smoothed over a 30 second period.

The 5 minute smoothing algorithm still reduced throughput but only by about 1%. At the same time it reduced end of queue conflicts by 30% and lane changing conflicts by 20%. Speed variances remained low. And they saw medium to high compliance with these advisory signs.

Compliance is key, of course, and the Missouri experience was very different from previous installations in places such as Utah and Oregon. Even where VSL posted limits were enforceable, compliance only occurred when message signs explained the reason for the speed limit reduction or when law enforcement was present and actively enforcing those limits.

The reasons for this could be many. Maybe Missouri drivers are just more law abiding. It was not discussed in the study but perhaps MoDOT did a better job of explaining VSLs to the public before they were installed. Or perhaps the need for variable limits is more apparent to drivers in a congested urban work zone. But whatever the reason, it is clear we should not give up on variable speed limit systems just yet. More studies are needed, especially on the subject of the best applications for these systems and the algorithms driving them.

Innovate.ATSSA.com

As most readers will know, I’ve been involved in work zone ITS for nearly 20 years now. So I assume most people are aware of the technology and aware of the availability of studies, best practices, specifications, and more. But one should never assume, especially in a discipline where new practitioners are arriving every day.

This was hammered home to me in a phone conversation yesterday. A fellow contractor complained about a state that just let a project with a work zone ITS spec that no one can meet. Another person on the call told a similar story about an engineering firm.

This is compounded by the fact that those not in our industry don’t know where to begin their research. For that reason the American Traffic Safety Services Association (ATSSA) has created the “go-to” website for work zone ITS. You will find it within their Innovate.ATSSA.com website at: http://innovate.atssa.com/work-zone-its.html

This website was just introduced to members in February. It is still new so we are adding resources every day. But because it is new, you can be sure that everything there is current and the best information available.

On the ATSSA website you can learn about new technology, you can search for projects by state, and you can view upcoming industry events where you’ll be able to learn more. There is a blog area where you can read this and many other work zone ITS – related blogs. And most important there is a large section devoted to news and resources. Check it out today!

Google Maps are Wrong!

At any given time, perhaps 5% of Google Maps data is wrong. And the reason is simple. Construction traffic control requires contractors to close lanes, redirect traffic into oncoming lanes, or close roads altogether until the work has been completed. Those closures are reported to state and local agencies. And those reports are picked up by Google and other traffic data aggregators. But they are often wrong or out of date.

In most states, contractors are required to request permission to close a lane. That request must be made well in advance of the date on which they wish to close the lane, 7 to 10 days on average. By the time that day comes long, construction delays, weather, and other issues often postpone the work and the lane closure does not take place.

Contractors also often make several requests so they will have a multi-day window in which they can perform the work. The days they don’t work are called ghost closures. Some states have moved to eliminate ghost closures by requiring contractors to call the local traffic management center when the lane is taken and again when it is opened back up. This certainly helps, but it does not eliminate the problem altogether.

To make matters worse, many closures are never reported at all. Utility companies are notorious for closing lanes without permission. They reason that they are only there for a short time and so won’t affect traffic all that much. But as traffic becomes more dependent on accurate travel time and route information, any disruption causes problems, and may even be dangerous.

Incident response closes lanes; school crossing guards stop traffic; special events close roads and reroute traffic; flooding, fires and other environmental events also result in route closures and restrictions.

This is an important point of discussion in the automated/autonomous vehicle world, too. If autonomous vehicles depend on historic GPS data to plan and drive a route, they will run into unexpected construction. So they must decide how they will adapt to changes in geometry, in the number and location of lanes, and much more. And delays resulting from these closed lanes and detours should be measured and included in any travel time algorithms.

It is worth noting that the folks in the traffic data companies know of the problem but they can’t solve it on their own. Industry is beginning to fill this need. Arrow boards and flagger stop/slow paddles are being reinvented to become “smart devices”. They report in automatically when work begins and ends. And they also report their precise location. As the work moves, that is reported as well, so map data for work zones can now be reported in real time.

Much work remains to be done. But the solution to this problem is clear. The closures must be reported in real time from the field. And that includes any changes in geometry when lanes are temporarily shifted in one direction or another. Highway construction, incident response and special events all experience unexpected changes on a daily and often hourly basis. Maps must reflect those changes if our system is to be as safe and efficient as possible.

Variable Speed Limit Webinar

Variable speed limits systems have always, at least intuitively, promised benefits for work zones including greater throughput, reduced speed variance, and as a result, fewer crashes. We discussed these systems in a post in October 2015 after a presentation at the National Rural ITS meeting in Utah. The concept made sense and we looked forward to greater use of VSL systems.

A webinar was just offered April 4th by the US DOT Office of Assistant Secretary for Research & Technology entitled, “Variable Speed Limit Systems – Are They For Everyone?” The speakers, and there were several of them, did a great job of explaining the advantages and disadvantages of these systems. Those speakers included Jimmy Chu of FHWA, John McClellan of MnDOT, Bryan Katz of Toxcel, Jiaqi Ma of Leidos, and Vinh Dang of WsDOT.

There wasn’t a lot of new information. Instead they presented a comprehensive history of VSL systems from around the country. They looked at different uses for these systems, including work zones. In all, more than a dozen different projects from 9 different states were discussed. Some projects were relatively small, and others like those in Minneapolis and Seattle, were quite large.

But in almost every case, the results of these systems have not met expectations.  This was true for weather related systems, work zones, and congestion management applications. Reductions in speed variance and crashes were very small if not non-existent. And the reason in every case was a lack of conformance by the traveling public. There was some smoothing, but very little.

Many of the system designers anticipated this and included variable message signs alongside the VSL signs to explain why the speed reduction was justified. But drivers either misunderstood when they were supposed to slow or simply chose to continue at their current speed until they saw the problem for themselves.

Law enforcement is critical for these installations. Without enforcement, compliance will never reach levels that will result in the benefits designers expected. But law enforcement often became distrustful of the data they received, or didn’t get timely notifications at all. They also ran into serious resistance from the courts. So enforcement slowed, and compliance tanked.

There is still hope that these issues will one day be resolved. But for now, variable speed limit systems just aren’t providing the benefits we all hoped to see.  The webinar closed with a short discussion of future considerations. One thought was to combine these systems with a larger big data process (as discussed in our last post). They might look at not just weather, or work zone conditions, but also at traffic speed and volume data approaching the area, timing of major events, and more to improve drivers trust of VSLs.

Another thought was with regard to automated vehicles. Will VSL systems be more effective when the information is sent directly to each vehicle? If a pop-up display recommends slowing to 35, will they be more likely to do so? Or will they continue to ignore them as they apparently do now? Once autonomous (driverless) cars are on the road, the recommendations from these systems will be adopted automatically. But until then, compliance will remain the biggest problem for VSLs.

Machine Learning and Work Zones

Last July we talked about the phenomenon of Poke’mon Go and augmented reality and considered ways in which that new technology might be applied to work zones. Today let’s consider another new technology. This one is not really new but is enjoying renewed interest thanks to Big Data and an abundance of inexpensive computing power. Using machine learning we can now process our data in ways we could not before.

In its simplest form, machine learning is an algorithm or model that goes through data looking for patterns. Then using those patterns it can make predictions for another, similar set of data. Machine learning works well with different types of data. So, for example, it could crunch through traffic speed data and compare it to social media sites focused on local traffic problems. In this way it can find trends or commonalities that don’t come through in conventional data analysis.

Machine learning is a faster and more accurate way of adapting to human factors issues related to work zones. We set up a system before the project starts and configure it for conditions at that time. But conditions change in work zones. So what worked early may not work as well later. Now don’t misunderstand. The system still works. It still warns of queuing ahead. But if the timing of the messages varied depending on changing conditions throughout the day, coming on sooner, or turning off faster, it could improve driver compliance. And the only way to learn that is through machine learning.

Now there is one problem with machine learning, and it is a big problem. It acts as a black box taking in data and spitting out recommendations. But the process is not something you can review after the fact to understand what influenced the outcome. So agencies could not tell their governor why they have made changes. But if you are more interested in being able to adapt to changing conditions quickly and accurately than in being able to explain it, machine learning is the tool for you.

Let’s consider a few possible applications:

  • Algorithms used to trigger messages could learn to de-emphasize input at certain sensors at certain times of the day when that sensors data is not in line with conditions elsewhere.
  • Models could make better recommendations about the best times and days of the week for lane closures while still meeting goals for minimal level of service.
  • In time, machine learning could establish baselines for crashes, fatalities, etc. resulting from a project both with and without work zone ITS systems in place. The system would be justifying itself! Of course, the predictions would have to prove true over time.
  • Predictive analytics reviewing probe and spot speed data may see locations of higher crash frequency or recommend a different geometry or staging in that location.

Is this complicated? Absolutely. But tools are already available to make it easy to get started, to choose the best algorithm, and to scale up and down in size as the situation demands. And Iowa DOT has already doing it at their Center for Transportation Research and Education.

Until now we could only design a system for a project to get the best overall performance. Now we can save even more lives through machine learning. Once our systems begin to learn from their own data, they will perform even better than they are today.

Sending Work Zone Warnings to Cell Phones

workzone-alertThe Minnesota Department of Transportation just released a study looking at one way of triggering in-vehicle messages in vehicles approaching work zones. The authors Chen-Fu Liao and Max Donath of the University of Minnesota tested the concept of sending low-energy BlueTooth messages to Android phones equipped with a custom app they call Workzone Alert.

The app triggers an audible, visible and/or tactile warning to the driver as he or she approaches a work zone. Drivers who are speeding can also be warned to slow down. And the app can even disable calling and texting while within the confines of the work zone.

Report 2016-38 entitled “Investigating the Effectiveness of Using BlueTooth Low-Energy Technology to Trigger In-Vehicle Messages in Work Zones” was published by the Minnesota DOT. You can download a copy HERE.

 

Their design worked well and proved that vehicles traveling at speeds of up to 70 MPH could receive warning messages as they approached a work zone.

We have talked about a future with DSRC “pods” transmitting to vehicles. We have also talked about those same DSRC devices attached to PCMS as a stop gap to reach all vehicles until nearly all are equipped with receivers. But a better way might be through 5G cell service as that is already available and in most vehicles.

These “BLE tags” use very little power so they could be attached to message signs or arrow boards without affecting the signs performance. When packaged with a small battery they could also be attached to a simple sign post or overpass.

The downside is the cell phones must currently be placed into a BlueTooth discover mode to find existing tags. This uses more power and results in reductions in charge life for the phones. But if this technology continues to show promise, the Android and iPhone operating systems could surely be changed to receive these messages in something similar to a discover mode but one that uses far less power when not receiving. The BLE tag locations are stored allowing phones to run Workzone Alert in background except when passing known tag locations.

They also attempted to make the technology easy to deploy. A second app was developed to make it easy for traffic control contractors to update the message that Workzone Alert displays for a specific BLE tag.

Work at the U of M continues. The current, second phase of research is looking at human-factors considerations for alerts. What wording and format should be used to get the best results? In the third phase they will look at how best to maintain the BLE tag database, who should be able to make changes, and if it is practical to tie this into 511, Waze or Google Maps. Stay tuned as this promises to develop quickly!

Enterprise to Evaluate Use of Arrow Boards for Real-Time Traffic Updates

enterpriseThe American Traffic Safety Services Association (ATSSA) just concluded their annual Traffic Expo held this year in Phoenix, Arizona. Their Innovation Council met on Saturday, February 11th. The meeting was “standing room only” and included several great presentations. One of them was by Dean Deeter, President of Athey Creek Consultants on a project they are leading for Enterprise. You may remember that Enterprise is a consortium of 14 states conducting pooled-fund studies to jump start promising new technologies. They like to do more than just study a problem or technology. Instead, their goal is often to develop an operational concept and system requirements for a promising new idea.

In this project they hope to develop an automated system to update traveler information systems with work zone conditions as they change. Their concept begins with an arrow board. But the arrow board would be equipped with a unit that is GPS enabled. When the arrow board is turned on, it would notify the traffic management center (TMC) over a digital wireless link. It would also tell the TMC what mode it is flashing (right arrow, left arrow, or caution mode). With that information the system will know when a lane closure begins, and which shoulder, lane or lanes are being closed. When the arrow board is turned off, it would notify the TMC that the closure has been removed.

In this way the system will provide specific, real-time information about each work zone. We aren’t doing that very often now. Instead, most work zone warnings are generic. Our portable message signs just say there is road work ahead and either to expect delays or use caution. That’s not as helpful as it could be.

Furthermore, travel websites like Google Maps only tell you there is a work zone. It may tell you when the lane closure is planned to begin and end. But that’s about it. Users cannot normally tell if that work zone will delay them enough to justify taking an alternate route.

The Enterprise study will develop this concept in phase one. In phase two it will work with one or more member agencies to integrate such a system into their ATMS, control permanent message signs, and more. Work Zone ITS Blog will continue to follow this and report developments as they are made available.

It is interesting that they have chosen to integrate directly with state ATMS systems. Many states’ IT security prohibit outside data sources. Only data collected from DOT sensors is used. That’s fine for permanent ITS but it is a real problem for the portable elements found in work zone ITS systems. States that operate within a closed network can never take full advantage of work zone data. So we hope they succeed. But only time will tell if they do, or if, instead, work zone data finds it’s way directly to end users through phone apps like Waze.