Adaptive traffic control solutions (ATCS) are an artificial intelligence product enabling traffic signals to adapt to actual, real-time demand rather than sticking to a fixed schedule. By giving vehicles the green light when and where necessary, “smart” signals can provide a 40-to-1 return on investment through reduced travel times, fuel usage, emissions, and accidents.

Many products claim adaptive signal control functionality, but broadly defined the technology is “intelligence that dynamically and quickly adjusts the behavior of one traffic signal or a network of traffic signals based on observed conditions.” This means the signal changes how much green light time is given to each approach, how often each approach gets the green signal, and the relationship between the green signals at adjacent intersections.

Solution components/cost factors

A product can be as simple as a piece of software or as extensive as a bundled package of controllers, processors, servers, and detection devices. Vendors combine the following elements in a number of ways.

Vehicle detection methodology. This is the greatest point of differentiation and has the most impact on cost.

Find out what type (inductive loop, video, magnetometer, radar) and how many (stop-bar, advance, mid-block, departure; critical and noncritical movements) detectors are necessary for optimal performance and whether the solution includes some or all of the required detection. Look for one that comes with detectors or works with your operation's existing detection methods.

Compatibility with existing resources. The more a solution integrates with your operation's signal cabinets and controllers, detection devices, communications, and central software, the less you pay.

Level of sophistication. “Complex” doesn't necessarily equal “smart.” Avoid hype and ask about the model of the adaptive algorithm. What does the model try to accomplish? How does it make those decisions?

Communication network type, bandwidth, protocol, and reliability. Adaptive solutions don't include the critical component that enables signals to “talk” to each other and the traffic control center. Most can operate on copper, wireless, or fiber-optic networks. Some can use serial communications; others require Ethernet protocol. Some include robust mechanisms to handle the interruption of communications; others have little or no backup plan.

Whether the solution relies on centralized or distributed architecture. Most use a centralized server located at the traffic control center to store data and make decisions. Solutions that make decisions at the traffic cabinet — what's known as distributed architecture — are less vulnerable to a communications failure.

How much control you can have, and how much input you must give. You don't want a black-box solution that you can't control, but you also don't want a system that needs babysitting. Some solutions require the operator to create, load, and maintain a library of timing plans the system chooses from and then downloads based on criteria; and an agency (or its consultant) can spend considerable resources building and maintaining these new timing plans.

Other solutions don't use timing plans and simply adapt to traffic dynamically.

Some solutions require active operation, while others are more set-and-forget.