As a typical motorist who hates sitting at long red lights, it is reassuring to know that there is so much that occurs behind the scenes to control traffic congestion. As we learned in the first blog (7 Traffic Myths Debunked), there are a number of reasons we get into a jam!
Here’s a ‘crash course’ in traffic engineering which will illuminate the real reasons traffic signals seldom function as we want them to.
Essentially, traffic signals can be considered to have 3 major parts: 1.) Traffic signal heads, 2.) Vehicle sensors or detectors, and 3.) Traffic signal controllers and cabinets. As with a police officer directing traffic in a busy intersection, these parts must work together to make the best decisions for optimal traffic flow. Thinking of the officer analogy, this makes sense: his brain would be the signal controller; his eyes the sensor; and his hands the signal head. He must effectively observe, make decisions regarding signals, and communicate them to motorists to ensure their safe and efficient travel; this is the same way traffic signals work.
1. Traffic Signal Heads
As pedestrians and motorists, we’re all extremely aware of signal heads; they are generally mounted on mast-arms or hung on wires spanning the intersection (aptly called “span wires”) and have three color-coded indications. As many of us have known since playing rousing games of “Red Light Green Light” as youngsters, these lights let us know when to stop and go, respectively. The middle, amber—or more commonly “yellow”—light alerts us to prepare to stop.
Most of us only think of this traffic signal component when expressing our frustration with red lights. When we must abruptly stop at yet another intersection, we’re more likely to shake our fists at that stubborn red light than any other part of the traffic signal system. In addition, these are often the most updated component of the three-part system. In fact, most signal heads now use Light Emitting Diode (LED) lights in order to conserve energy. While signal heads may have the greatest presence in the public view of traffic signals, they are not as critical in signal optimization as the other two components.
2. Vehicle Sensors
The next most infamous component is the subject of many of the myths described in the first blog: the vehicle sensor. The vehicle sensor detects the presence of vehicles at the stop bar of each lane, though it is neither a laser curtain nor a buried scale. The information collected by the sensor is binary, reading either “yes, at least one vehicle is present” or “no, not even one vehicle is present.”
Depending on the technology utilized, a sensor may or may not recognize the presence of more than one car. Once collected, this information is sent to the traffic signal controller for processing. The sensor is usually an inductive loop located beneath the pavement, though newer video technologies are growing in popularity. Two other less popular methods of detection include radar-based detection and detection using wireless magnetometers. Regardless of its form, the role of the vehicle sensor is always to detect cars and communicate that information to the signal controller.
…Role of the vehicle sensor is always to detect cars and communicate that information to the signal controller.
Most intersections currently employ inductive loop sensors. Electromagnetic induction-based detectors use coils of wire under each lane. These coils are commonly referred to as “inductive loop detectors” or “loops.” This method works based on the laws of electromagnetic induction; when a mass of metal is moved inside a coil of wire, electricity or an electric pulse is generated. With these sensors, the vehicle is the mass of metal that moves inside the coil of wire buried under the lane. The detector constantly monitors the electric pulse in the loop and informs the traffic signal controller about the presence or absence of a car. Since induction-based detection relies directly on the laws of physics, if configured properly, it offers the most reliability.
While inductive loop sensors provide reliable feedback, installation and maintenance are costly and time consuming. To install loops, the pavement on each lane is cut with a saw in the shape of the loop. Loop wires are then installed into the saw cut and sealed. Over time, pavement moves and as a result, loop wires are broken. Regular pavement maintenance projects also destroy loops, necessitating premature reinstallation.
Furthermore, this methodology is limited: it does not know whether there is one vehicle or a string of vehicles waiting. Loop sensors can only detect the presence of cars idling at the stop bar. As a consequence, loop sensors do not allow for more complex traffic mitigation, which could give preference to busier streets in order to relieve traffic congestion. If you have seen a solitary car on the side street given priority (and green indication) over a long line of cars approaching on the main street, this limitation is likely the reason.
Another detection method is based on video image processing. This method has gained popularity because its installation is less intrusive and its maintenance is easier. In video based detection, cameras are mounted on mast arms from which vantage point they record approaching traffic.
These (video) sensors are the little white, tubular objects sticking up from the signal mast-arm and staring at you. Fear not, these sensors aren’t recording your speed or your steering wheel drum solo!
The cameras send real-time images to an image processor. When pixels in the image don’t change over time, the processor saves them as a “learned” background image. The processor compares the real-time image with the learned image by literally subtracting the real-time image from the learned background image. If objects are left over from this subtraction process, the processor reports that a new car is present in its detection zone.
Installing cameras on mast arms is the only invasive step for deploying video image processing detection. If a camera fails, unlike replacing the loop, where the pavement has to be saw cut again, it is relatively easy to replace the failed part. However, video detection is susceptible to camera issues such as being blinded by the sun, fog, and snow. Additionally, shadows of moving objects and reflected light from wet pavements can generate false detection triggers. These false detection triggers bring up green lights even when there are no cars waiting to be served. Fortunately, software exists to readily compensate for these issues, thereby eliminating these issues and optimizing video detection sensors.
3. Traffic Signal Controller/Cabinet
If you’ve ever wondered about the purpose of the metal box at the corner of every major intersection, aside from being a public bulletin board or convenient place to unload an unwanted wad of gum, it is actually the traffic signal cabinet. The cabinet houses the signal controller and other hardware required to run the traffic signal. Based on input from the sensors set up at every intersection and the programming done by a traffic engineer, the signal controller determines the amount of green or “go” time given to each movement of cars. As such, it functions as the brains of the operation—aside from the traffic engineer, of course.
While they may look similar from the outside (i.e., big, metal boxes, possibly covered with graffiti), not all traffic signal controllers run in the same fashion. In fact, there are three modes of signal operations in the popular analog traffic signal controller. Most traffic signals operate using one of the following modes: pre-timed, actuated or synchronized.
The simplest of the three modes is the pre-timed mode. In the pre-timed mode, each vehicle movement is given a predetermined amount of time regardless of vehicle demand. Vehicle sensors are not required at these intersections because the system runs on a set timer rather than responding to the presence or absence of cars at the stop bar. Synchronization between signals is possible with the pre-timed mode; however, this method is very inefficient in terms of mitigating traffic or serving the needs of commuters and is not recommended.
Coming in as the polar opposite of the pre-timed mode is the actuated mode. In the actuated or “free” mode, there are no timing plans or synchronization between lights.
Each signal gets input from its own sensors and changes signal indication (i.e., the color of the light on the signal head) in a sequential fashion, regardless of what other signals on this main route are doing.
That is, with actuated signals, each intersection is something of a “first come, first served” affair: green lights and red lights are distributed according to the presence or absence of cars detected by the sensors at each individual intersection.
The advantage of the second mode, actuated mode, is that delay—that pesky wait time—on the side street is minimal during periods when traffic flow is light. The disadvantage is that delay experienced at the intersection as a whole will increase as motorists on the main street will need to stop while vehicles on the side street are being served. Also, there will not be any synchronization between signals. This lack of coordination will cause a substantial increase in stops, vehicle emission, delay, and fuel consumption. Thus, while isolated signals may effectively operate in actuated mode, it is not efficient to operate a string of signals in actuated mode.
The final and most efficient option is the synchronized mode. In this mode, the signal rests in green for the main street unless vehicles are present in the side street. This design takes into account the greater volume of traffic travelling through main streets and distributes more “go” time to that thoroughfare. Often, agencies may not install sensors on the main street because the signal will always revert to the default of giving green time to vehicles travelling on that street. The assumption is that more vehicles will always be present on the busier corridor.
Additionally, it is possible to synchronize multiple signals in an arterial in this mode, thereby allowing even greater customization for the needs of commuters along that route.
The Holy Grail every traffic engineer seeks is an efficiently installed and maintained system that allows for flexibility depending on changes in traffic patterns; thus, the synchronized traffic signal is the preferred option for effectively mitigating traffic congestion in city driving conditions.
Synchronizing semi-actuated signals allows whole movements of vehicles along a specific route to travel in what are called “green tunnels.” A green tunnel is a passageway of green lights enabling a group, or “platoon,” of vehicles to travel uninterrupted by congestion-causing red lights. Moreover, all signals that are synchronized must have the same cycle length. The cycle length is defined as the time it takes the timer within the controller to turn 360 degrees. In terms of what you see as a motorist, a cycle length is the time it takes to move from the beginning of green light to the end of a red light for your main street. Within that duration, the controller is required to sequentially serve all movements that have a demand. While this information may seem technical, we’ve all seen it in action. In reality, where an intersection fails to serve all movements sequentially—that is, give each direction of traffic its share of green lights—the resulting chorus of car horns would alert you of the lapse.
The cycle is further divided into “splits,” where splits are the percentage of a cycle allocated to a particular vehicle movement. Think of it this way: at an intersection of a main thoroughfare and a smaller side street, the main road is going to get the larger “split” of the cycle or the larger cut of green time. There is also one fixed point in the cycle that can be considered as a baseline and referenced. Offsetting this fixed point from one intersection to another synchronizes signals. If you’ve ever sung a tune like “Row, Row, Row Your Boat” in rounds, you can imagine the first singer as establishing the baseline upon which the other singers find their rhythm and melody. Much like when singing in rounds, synchronizing traffic signals in this fashion allows for a more complex, harmonious flow of traffic.
How Signals Communicate?
To delve even deeper, traffic signals should be interconnected to form a network to ensure optimal performance. Since roadways already function as a network, it makes sense for the technology responsible for directing traffic on these roads to communicate the same way. Traffic issues don’t occur in isolation; if they did, we wouldn’t need to listen to the traffic reports during our morning or evening commutes.
Traffic runs best when signals are interconnected and work together in a synchronized manner. Common media used for interconnection include wireless, fiber-optic cable, and twisted pair copper conductors. Often, this communication network covers the entire city and terminates at the office of the traffic engineer. Interconnection provides the engineer with remote access to the signal but synchronization is not perfect. Nevertheless, the overall effect of networking signals maximizes the efficiency of signal performance.
We hope you’ve begun to understand some of the technical reasons behind the difficulty of optimizing traffic signals. The foundational equation behind most of the software models is flawed, and furthermore, the process of synchronizing the signals based upon those flawed equations is itself a time consuming and expensive endeavor.
Since synchronizing traffic signals in this traditional manner is so inefficient, between 70 and 90 percent of traffic signals in the United States are not synchronized. In fact, the National Transportation Operation Coalition grades U.S. traffic signal operation a grade D.