1.4 Central Traffic Management Systems (ATMS) & Architecture

1.4 Central Traffic Management Systems (ATMS) & Architecture

Modern traffic signal networks do not operate in isolation. They are managed through central software platforms known as Advanced Traffic Management Systems (ATMS). An ATMS integrates field traffic signal controllers, dynamic message signs (DMS), CCTV surveillance cameras, automated incident detection systems, and connected vehicle infrastructure into a unified monitoring and control environment.

---

1. ATMS Deployment Architectures

ATMS deployments fall into three architectural categories, each presenting distinct trade-offs in communication bandwidth, computing topology, and fault tolerance:

CENTRALIZED ARCHITECTURE          DISTRIBUTED (EDGE) ARCHITECTURE      HYBRID CLOSED-LOOP
┌──────────────────────┐          ┌──────────────────────┐             ┌───────────────────┐
│ Central TMC Server   │          │ Central TMC Server   │             │ Central Workstation│
│ (All Logic & Storage)│          │ (Monitoring / Config)│             └─────────┬─────────┘
└──────────┬───────────┘          └──────────┬───────────┘                       │ IP Backbone
           │ Cont. 1-Hz                      │ Supervisory Status                ▼
   ┌───────┴───────┐                         ▼                         ┌───────────────────┐
   │ High-Speed IP │              ┌──────────────────────┐             │ Arterial Master   │
   └───────┬───────┘              │ Edge IP Network      │             │ (Subsystem Logic) │
           │                      └──┬────────┬────────┬─┘             └─────────┬─────────┘
   ┌───────┼───────┐                 │        │        │                         │ Field Bus
   ▼       ▼       ▼                 ▼        ▼        ▼                         ▼
┌─────┐ ┌─────┐ ┌─────┐           ┌─────┐  ┌─────┐  ┌─────┐                   ┌─────┐ ┌─────┐
│Local│ │Local│ │Local│           │Edge │  │Edge │  │Edge │                   │Local│ │Local│
│Ctrl │ │Ctrl │ │Ctrl │           │Ctrl │  │Ctrl │  │Ctrl │                   │Ctrl │ │Ctrl │
└─────┘ └─────┘ └─────┘           └─────┘  └─────┘  └─────┘                   └─────┘ └─────┘
 (Synchronous Remotes)              (Peer-to-Peer Interconnect)                 (Secondary Units)

A. Centralized Architecture

• Mechanics: All signal timing optimization, coordination scheduling, split calculations, and database archives reside on enterprise servers at the Traffic Management Center (TMC). Field controllers operate as synchronous remotes, receiving phase hold, force-off, and cycle commands every second over a high-speed fiber-optic or wireless IP network.
• Advantages: Massive computing power, centralized database management, simple controller firmware, and seamless wide-area network optimization.
• Disadvantages: Extreme vulnerability to communications disruption. If the central server or wide-area network (WAN) fails, local intersections cannot adapt and must drop into standalone time-base coordination.

B. Distributed / Edge Computing Architecture

• Mechanics: Optimization and control intelligence reside directly at the intersection within Advanced Transportation Controllers (ATC) or auxiliary cabinet edge computers. Intersections share speed, volume, and platoon trajectories peer-to-peer across an industrial Ethernet network.
• Advantages: Superior system survivability. A complete severance of the central TMC link causes zero interruption in localized adaptive coordination.
• Disadvantages: Increased cabinet hardware costs, more complex field configuration management, and decentralized firmware updating requirements.

C. Hybrid Closed-Loop Systems

• Mechanics: A central workstation communicates periodically with designated field Arterial Master Controllers, which in turn manage groups of secondary intersection controllers along a corridor over local serial (RS-485 / FSK) or Ethernet sub-buses.
• Advantages: Balances central visibility with corridor autonomy; ideal for agencies with legacy field infrastructure.

---

2. The NTCIP Standards Suite for Signal Control

Prior to the mid-1990s, traffic signal communications relied on closed, proprietary vendor protocols. An agency purchasing Controller Brand A was locked into purchasing System Software Brand A. To eliminate vendor lock-in, the National Transportation Communications for ITS Protocol (NTCIP) was developed through a joint effort of AASHTO, ITE, and NEMA, sponsored by the FHWA.

Core NTCIP Standards in Signal Systems

NTCIP 1201: Global Object Definitions

Defines standardized Management Information Base (MIB) objects common to all ITS devices:

• System Clock & Timekeeping: Real-time clock objects, Daylight Saving Time adjustment rules, leap second adjustments.
• Device Configuration: Unit identification, firmware version tracking, memory management, and warm/cold reset commands.
• Event Logging: Event configuration tables, event logging thresholds, and circular event-buffer definitions.
• Time-Base Scheduling: Day plans, week plans, and holiday schedule calendars.

NTCIP 1202: Actuated Traffic Signal Controller Units (ASC)

The definitive standard for traffic signal control. NTCIP 1202 defines standard MIB objects governing all internal and external controller operations:

• Phase Objects: Minimum green, maximum green 1/2, yellow change, red clearance, vehicle extension (passage), red revert, walk, and flashing don't walk.
• Ring-Barrier Configurations: Dual-ring concurrency, barrier definitions, and phase sequencing assignments.
• Detector Objects: Real-time detector call status, detector mode (pulse/presence), and volume/occupancy reporting.
• Coordination Objects: Coordination pattern tables, cycle length, split values, offset values, sync source selection, and coordination mode (permissive, yield, force-off).
• Real-Time Override Objects: Remote phase hold (phaseStatusGroupHold), remote force-off (phaseStatusGroupForceOff), phase omit (phaseStatusGroupOmit), and vehicle/pedestrian call injection.
• Preemption & Priority Objects: Emergency vehicle preemption calls, railroad preemption routines, transit priority calls, and preempt dwell/exit phase tracking.

NTCIP 1103: Transportation Management Protocols (TMP)

Defines network and application layer protocols for bandwidth-constrained Center-to-Field links:

• SNMP (Simple Network Management Protocol): Standard get/set/trap protocol operating over UDP/IP for device configuration, status polling, and alarm delivery.
• STMP (Simple Transportation Management Protocol): A specialized dynamic protocol that groups disparate MIB objects into compact binary packets, reducing communication overhead across low-bandwidth wireless links.

NTCIP 2306: Application Profile for XML Message Encoding (C2C)

Defines message encoding using the Extensible Markup Language (XML) and Simple Object Access Protocol (SOAP) over standard HTTP/HTTPS for Center-to-Center (C2C) communications between distinct software platforms.

---

3. Center-to-Field (C2F) vs. Center-to-Center (C2C) Communications

---

4. Master Clock Time Synchronization & Drift Tolerances

In coordinated arterial signal systems, the spatial relationship between signal indications is governed entirely by the Coordination Offset. A signal's offset is referenced to a master time standard.

The Physics of Time Drift on Arterial Progression

Consider an arterial corridor with signals spaced at $2,000\text{ feet}$, an operating cycle length of $C = 100\text{ seconds}$, and a design progression speed of $45\text{ mph}$ ($66\text{ ft/s}$):

• The physical travel time between intersections is: $t_{\text{travel}} = \frac{2,000\text{ ft}}{66\text{ ft/s}} = 30.3\text{ seconds}$
• The downstream green offset must open exactly $30\text{ seconds}$ after the upstream green.
• If the downstream controller's internal clock drifts by $+5\text{ seconds}$, its green window opens 5 seconds late relative to arriving traffic.
• In physical distance, a 5-second offset error shifts the progression band by: $\Delta d = 5\text{ s} \times 66\text{ ft/s} = 330\text{ feet}$
• The leading 330 feet of the platoon (approximately 12 to 15 vehicles) slams into a red indication, destroying progression, causing unwarranted stops, and precipitating rear-end collision hazards.

Synchronization Protocols & Engineering Tolerances

• NTP (Network Time Protocol / RFC 5905): Local controllers poll Stratum-1 or Stratum-2 NTP servers over the IP network, synchronizing clocks within $\pm 10$ to $50\text{ milliseconds}$.
• GPS Time-Sync Cards: Cabinet-mounted GPS receivers supply a physical Pulse-Per-Second (1PPS) hardware interrupt and NMEA-0183 time sentences directly into the controller, maintaining sub-millisecond synchronization independently of network availability.
• Drift Tolerances:
  • $<100\text{ milliseconds}$: Mandated tolerance for closely spaced, tightly coordinated ASCT and high-speed arterial systems.
  • $\pm 500\text{ milliseconds}$: Maximum acceptable drift for general urban coordinated networks.
  • $>2.0\text{ seconds}$: Unacceptable operational failure. At this threshold, controllers initiate disruptive transition cycles (dwell, smooth, or add/subtract) to realign, severely fracturing arterial flow.

---

5. High-Resolution Event Logging & Diagnostic Dashboards

Modern traffic management leverages Automated Traffic Signal Performance Measures (ATSPMs) based on high-resolution event data (the Indiana 10-Hz standard):

High-Resolution Event Data (10-Hz Logging)

The controller records every internal state transition with a timestamp accurate to one-tenth of a second (100 ms):

• Event Code 1: Phase On (Green display begins)
• Event Code 8: Phase Yellow Clearance Begins
• Event Code 9: Phase Red Clearance Begins
• Event Code 21: Pedestrian Walk Begins
• Event Code 22: Pedestrian Flashing Don't Walk Begins
• Event Code 81: Detector Actuation (Vehicle enters detection zone)
• Event Code 82: Detector Release (Vehicle clears detection zone)
• Event Code 102: Preemption Input Activated

Key ATMS Diagnostic Displays

• Split Monitor Log: Tracks actual phase green durations cycle-by-cycle against programmed splits, highlighting whether phases terminate via Gap-Out (demand satisfied), Max-Out (unmet demand remaining), or Force-Off (coordination boundary reached).
• Purdue Coordination Diagram (PCD): Plots detector actuations relative to cycle time and green intervals, providing visual proof of whether platoons arrive on green or red.
• Cabinet Alarms: Real-time automated alerts notifying technicians of MMU/CMU flash trips, cabinet door openings, detector faults, power outages, and UPS battery backup transitions.