Adaptive intersection control gets a lot of buzz, but it can be expensive and difficult to manage. Knowing best practices can help make deploying an adaptive system a win-win situation.
In a past life I was a sports writer covering mostly high school athletics. These games were a blast to watch because the kids were new enough to the world of sports to be excited about everything, yet cocky enough to show o every now and then. I remember one game in particular—a rivalry basketball game between Cedar and Canyon View, the two high schools in Cedar City, Utah. Canyon View wasn’t doing so well. They were playing a man-to-man defense and the Cedar kids were running them ragged. They would switch out who was covering whom, but Cedar kept them off balance. I especially remember two Canyon View kids slamming into each other as their offensive counterparts laughed at them. They were being led around like puppets.
Finally, the coach called a timeout and changed the strategy. They switched from a man-to-man defense to zone. Suddenly, it was like a different game — the Canyon View kids would assemble in formation and keep the Cedar kids in the far perimeter of the key; each time the offense got the ball there were two or three defensive players in their face.
Canyon View came away with the victory that day. Talking to one of their star players afer the game, he told me they needed their coach to call a new defense — that he saw what the players couldn’t see and that adaptability was what won the game.
Wavetronix research engineer Brad Giles likes to use basketball as an illustration for adaptive signal systems in the traffic world. If you look at three different levels of adaptability in traffic lights, the basketball analogy is clear: first, there are fixed systems where light cycles are run on a timer, like an electronic game of basketball where everything is decided by math and expected to run properly; second, there are actuated systems where splits and sequences for traffic lights can change if vehicles are detected — a side street may get a green light early, for example. This is like the players on the court trying to figure out what the other team is doing and responding in kind to their play style.
On the third level, you have flexible adaptive systems. This is where the coach sees the game is changing and dramatically alters the playstyle of their team. In traffic, a master controller is fed information about current traffic conditions and can change the system to meet the current need. is type of flexibility and adaptability in the traffic world is still largely unused. The benefits it can provide, however, justifies a closer look at how adaptive systems are used; the benefits and problems inherent in adaptive systems; and some best practices for adopting them.
In many ways, you can look at adaptive systems as a natural evolution in traffic signal operation. Early traffic signals operated on a timer independently from other traffic signals. As technology improved, especially regarding traffic detection and communication, signal operations improved as well.
Brett Sellers, an industry veteran with experience in the public and private sectors, talks about how early, and even some modern, systems entered the world of actuated systems.
“Signal timing plans were based off volume counts,” Sellers says. “Data could be collected by pneumatic tubes or by placing someone at an intersection to record how many cars went in each direction.”
This data was analyzed and timing signals were created that matched the data. From this, traffic engineers could create time-of-day schemes, meaning intersections would behave differently during periods of heavy traffic. The data would, hopefully, be supplemented with a personal touch from engineers who understood how traffic in the area worked beyond the limited snapshots offered by times of data collection.
The limitations of this system are apparent. Though efficiency was increased, they reacted to historical time-of-day demands and not actual traffic. These time-of-day plans could be changed, but that required further data collection and extra time and budget from the agency to do so.
Adaptive systems take advantage of further advances in detection and communication technology. Rather than use temporary counting measures, permanent detection allows for 24-hours-a-day, seven- days-a-week data collection in real time. With the increased use of fiber optic and wireless communication, this data is available as it happens. The logical step is to use this real-time data and analysis to create systems that adapt to real-time traffic conditions.
There are dozens of adaptive systems on the market with many more one-off systems developed for specific locations and situations, making it difficult to summarize how adaptive systems work. Many use proprietary technologies and processes that are unique to a brand, but there are a few generalizations that can be made.
As stated by the Federal Highway Administration’s Center for Accelerating Innovation, adaptive system control technology works by following three steps: first, it collects data on current traffic conditions; second, it analyzes the data and evaluates possible plans to increase efficiency using system algorithms; third, it implements changes to the traffic control system based on the outcome of the evaluation.
Within this outline, specific tools and processes must be in place to make adaptive systems function. First is the need for quality detection. Without accurate, dependable vehicle detection, the quality of the data suffers along with every subsequent step.
There is also the need for reliable communication. This is usually achieved through a closed-loop system. ese systems require the ability for the individual intersections to communicate back and forth with a master controller, meaning there needs to be a sophisticated communication board in each local controller. A closed-loop system is also resistant to accidents. If someone was to sever a cable, for example, an open-loop system would fall apart, but a closed-loop system could still mostly function. ese systems are also easy to monitor remotely via modem. Closed-loop systems are only possible with the higher bandwidth available through fiber optic or spread spectrum radio communication systems.
With the data collected on reliable detection devices and reliably communicated, the analysis needs to be high performance to see true adaptive advantages. This is where things segment between different adaptive systems. Generally, systems use a traffic model to evaluate
the traffic data. These models can focus on individual intersections or entire systems. They can analyze data on a second-by-second basis or in a cycle. Likewise, the parameters used in a traffic model can vary in number and complication from several thousand parameters to a handful. Again, the Center for Accelerating Innovation states that the accuracy of a mode is improved when more parameters are considered, though it can be a time-consuming and expert-driven process to adjust and calibrate these models. Often the human touch is needed, according to Sellers.
“The traffic engineer has to know what they’re doing because the best computer in the world is still just a computer,” Sellers said.
It’s easy to look at the history of adaptive systems and say it’s a relatively new thing, but the truth is they’ve been around for nearly 30 years, though less than five percent of existing traffic signals utilize them. When looking at why we haven’t seen wider usage of adaptive systems, it’s important to consider a few points.
First, when using public funds, traffic agencies can be hesitant to try new things. When investing in new processes or technologies, it can be disastrous when a payoff doesn’t occur, and to be fair, there have been many examples in the traffic industry where new technologies and processes didn’t pan out. With such a limited use of adaptive systems currently in use, it’s easy to see why agencies may be hesitant to experiment with them.
Second, adaptive systems may not be right in every situation. Giles points to a couple of situations in Utah where the Utah Department of Transportation justifies using, and not using, adaptive systems to illustrate this point.
“In Salt Lake City, they have a pretty good idea what traffic is going to do and so their time-of-day plans work pretty well,” Giles says, adding that UDOT’s signal performance metric system allows them to make adjustments easily if needed.
Meanwhile, in Park City things are different. “In Park City, there are a lot of events and seasonal traffic that is hard to predict. There they use an adaptive system and it works well,” Giles says.
He adds that sometimes simply upgrading a signal’s detection and timing can show improvements without needing adaptive systems. To use the basketball metaphor again, if you know the other team isn’t going to try anything new, you already know how to win.
To Adapt or Not
Giles has a list of situations where adaptive systems can have the biggest positive impact on traffic, based both on improved performance and, sometimes just as important, methods to improve public perception and popularity of adaptive systems.
The first, and most important, feature is, again, a high-level of reliable detection and communication. Giles said this step should be the first priority, not only because it’s the base of the entire concept of adaptive systems, but it may show you don’t need an adaptive system in the first place.
“After updating detection, you may find you have already achieved a significant benefit in performance and may not need to spend the additional money for adaptive control,” Giles says.
When placed in an appropriate situation, however, adaptive systems can have a huge positive impact. FHWA says that adaptive systems will cut costs, both for agencies and travelers, increase customer satisfaction by limiting time spent in traffic, reduce emissions of hydrocarbons and carbon monoxide, and more. Since, again, less than five percent of existing traffic signals in the US use adaptive systems, the exact benefits are hard to gauge, but FHWA says studies show adaptive systems “improve average performance metrics (travel time, control delay, emissions, and fuel consumption) by 10 percent or more. In systems with particularly poor conditions, the improvement can be 50 percent or more.”
With these types of improvements possible, it’s worthwhile for departments of transportation to see how and where adaptive systems can be used in their network. It could be the difference between a team with a coach who can call the plays when they are needed, and a group of kids running into each other on the court.