Traffic signal systems: controllers, ATMS software, and cabinet communications

Reference article · 5 min read

A traffic signal system looks like one thing from the street, but it is four systems stacked on each other: the hardware in the cabinet at each intersection, the detection that tells the controller a vehicle or pedestrian is there, the communications path between the intersection and a central office, and the software engineers use to build timing and watch for trouble. Agencies buy these pieces at different times from different vendors, and trouble starts when no one owns the seams between them.

Ownership is often split. A city may own most signals on its streets while the state DOT owns those on state routes through town, and a county may maintain signals for smaller jurisdictions under agreement. Before any technology decision, write down for each intersection which agency owns it, which one maintains it, who holds timing authority, and who pays for the communications circuit. Those answers are often different agencies, and the agreements governing them outlive software contracts.

The field side: cabinets, controllers, and detection

• Cabinet. The cabinet standard, such as NEMA TS 1 or TS 2, a Caltrans-style 332, or an ATC cabinet, sets the wiring and load switch arrangement. Mixed cabinet types raise spare parts and training costs.
• Controller. The controller runs the signal program and is the real timing engine. Central software sends and reads parameters, but the controller keeps running when the network does not.
• Conflict monitor or malfunction management unit. An independent device that drops the intersection to flash on conflicting indications. It sits outside the software path by design, and its test records are a maintenance obligation.
• Detection. Loops, video, radar, magnetometers, and pedestrian push buttons. Many complaints filed as bad timing are really a failed detector.
• Preemption and priority. Railroad preemption is safety critical and typically hardwired. Emergency vehicle preemption and transit signal priority arrive through optical, GPS, or radio equipment and need written rules about who may request priority.

Central software: what an ATMS actually does

A central advanced traffic management system stores timing plan libraries, schedules them by time of day, pushes and verifies parameters, collects alarms and event logs, and often ingests high resolution data for performance measures. What it does not do is run the intersection. Design so an outage degrades gracefully: controllers fall back to internal schedules, coordination holds through time base sync, and staff know what the street will look like while the link is down.

NTCIP, particularly its object definitions for actuated signal controllers, makes a multi-vendor network possible, though vendors add extensions on top. Ask which demonstrated features depend on proprietary extensions, because those are the ones you lose if you change controller brands.

The communications path back to the shop

• Agency fiber. Highest capacity and no recurring circuit cost, but capital heavy, and it brings obligations: conduit records, splice documentation, and locate response.
• Wireless point to point or mesh. Useful for gaps and where trenching is impractical. Plan for line of sight, interference, and equipment aging.
• Cellular and leased circuits. Fast to deploy, with recurring cost per site. Insist on private APN or equivalent isolation, and confirm repair timelines, because signal outage response windows are usually shorter than standard telecom service levels.

Controller data is small; bandwidth planning is driven almost entirely by video, so decide early whether video belongs on the same path as signal control.

Cabinet network security and remote access

Cabinets are network equipment in unattended locations, physically accessible to several agencies and contractors.

• Put the signal network on its own segment with controlled routing to enterprise IT.
• Change default credentials on switches, radios, cameras, and controllers, and keep them in the agency password system, not a shared sheet.
• Give vendors brokered, time-bound, logged remote access with multifactor authentication, not a standing VPN account or modem.
• Define who may push firmware, and require a tested rollback.
• Alarm cabinet doors and reconcile openings against work orders.

Critical infrastructure security requirements change over time; confirm current expectations with the official source and your counsel.

Monitoring, alarms, and who responds

Define severity tiers before turning alarms on: dark signal, flash, conflict monitor trip, communications loss, detector failure, and cabinet door open. For each tier, write the response path, including after hours, the fallback when the central system itself is unavailable, and how the public safety answering point learns about a dark intersection. Tune thresholds and suppress known noise, because an alarm list nobody trusts is the same as no alarms.

Keeping timing plans and records out of one technician's head

• Store signed, versioned timing sheets in an agency system of record, not only in the central database or on one laptop.
• Keep cabinet prints, as-builts, loop locations, IP address plans, and fiber splice records where the next person can find them, and tie assets to the work order system so history follows the intersection.
• Require documented approval for timing changes and the ability to revert to the prior plan.
• Export the central database on a schedule and verify the export is usable without the vendor's application.

What to ask vendors and put in the contract

1. Request NTCIP conformance statements and a written list of functions that require proprietary extensions.
2. Ask what happens at the intersection during a central outage and a communications failure, then test both at acceptance.
3. Require data ownership, periodic export in a documented format, and clear end-of-contract terms.
4. Ask for patch cadence, vulnerability disclosure practice, remote access method, and which subcontractors touch the system.
5. Require training sufficient for agency staff to configure timing without the vendor, and acceptance testing at real intersections, including preemption and pedestrian operation.

HQT works with state and local agencies on systems engineering, integration, and cybersecurity for operational technology environments such as traffic signal networks.

Frequently asked questions