Office of Operations
21st Century Operations Using 21st Century Technologies

Integrated Corridor Management (ICM) Program: Major Achievements, Key Findings, and Outlook

Chapter 2. ICM Research Program Achievements and Findings

The ICM program began with two main objectives 1) to demonstrate and evaluate pro-active integrated approaches, strategies, and technologies for efficient, productive, and reliable operations; and 2) to provide the institutional guidance, operational capabilities, and ITS technical methods needed for effective Integrated Corridor Management. This chapter discusses how these objectives were met. The major phases of the USDOT ICM program are shown in Figure 1.

The USDOT initiated the ICM Program to research the integration of transportation networks within urban travel corridors. The first phase of the program was the foundational research phase, discussed briefly in Chapter 2, which involved research into the institutional, operational, and technical integration of individual corridor networks and development of ICM strategies to facilitate this integration; the goal was to see what would be needed in terms of cooperation, harmonization, and utility when formerly stove piped agencies would work in concert. Next, corridor tools and ICM strategies (e.g., ramp metering, congestion pricing, cross jurisdictional signal coordination and optimization, transit priority, and enhanced traveler information) were developed along with a framework created to model, simulate and analyze the strategies. After this initial research, the program solicited actual corridor stakeholders through a 3-stage competitive funding process to develop concepts for integrated operation of their corridor networks, analyze them to determine potential benefits, and then demonstrate them. Because these agencies were breaking new ground, their corridors were referred to as "Pioneer sites." The Pioneer site process is described later in this chapter. Briefly, Stage 1 began with the development of concepts of operations (ConOps) and system requirements in eight urban areas. Stage 2 involved analysis, modeling and simulation (AMS) of three corridor networks (Dallas, San Diego, and Minneapolis). Stage 3 consisted of awarding two sites full deployment grants; U.S.75 in Dallas, TX, and I-15 in San Diego, CA, thereby providing "proof of concept" demonstrations.

This graphic shows the 4 major phases of the U.S. DOT ICM research program, from left to right, the phases are: foundational research, Pioneer corridor sites, deployment planning, and mainstreaming ICM. In addition, ongoing knowledge transfer is shown as an activity for all project phases.
Figure 1. Chart. Major phases of the USDOT ICM research program.
(Source: USDOT)

Building off the successful Pioneer site process, USDOT initiated an ICM Deployment Planning Grant Program. The USDOT provided 13 sites up to $200,000 per site to enable those areas to develop "pre-implementation" documents (e.g., ConOps, Systems Engineering, and Project Management plans) and begin active planning for integrated corridor management systems, requiring 20% match by the local jurisdiction. This program also provided an opportunity for USDOT to test the effectiveness of the knowledge and technology transfer products (KTT) and activities that had been developed to date. The grant program results are discussed in Chapter 2 under "Promoting ICM Planning and Deployment."

As shown in Figure 1, knowledge transfer activities have been an ongoing part of every program phase. The final phase of the ICM research program is referred to as "Mainstreaming ICM" and consists of the continued knowledge transfer of ICM concepts, methods, tools, and products, to encourage the adoption of ICM into everyday transportation planning, project development, and operations. Mainstreaming activities also include focused research to assist with deployment challenges and policies to encourage deployment, such as establishing ICM as an eligible or even preferred project type in various deployment grant programs. More information on the original research plan can be found on USDOT's ICM website2. To learn more about ICM and view materials developed as part of the ICM Initiative, visit the ICM Knowledgebase3.

Foundational Research and ICM Concept

The original concept of ICM was defined in a white paper entitled "Conceptualizing Integrated Corridor Management". This paper is included as Appendix A in the "Integrated Corridor Management: Implementation Guide and Lessons Learned (February 2012)." 4 That historical document defines the original ICM concept of operations and identifies some early "startup" lessons learned. However, be advised that we intend to publish an updated companion executive summary called "Mainstreaming ICM: An Executive Level Primer" in late 2019 which intends to address the questions "how do we get started," and "how and why should our region invest in ICM," and, for that matter, "what constitutes a 'candidate' ICM region?" We strongly encourage startup regions to pair this document, which explains to executives what ICM is, with that one that is planned, which will explain how to adopt ICM in your region.

The ICM concept paper hypothesized or envisioned that managing a corridor in an integrated fashion requires corridor network operators to develop strategies in four areas and implement those strategies in one or more areas. The four areas include:

  • Demand Management
  • Load Balancing
  • Event Response
  • Capital Improvement

Within the first three strategic areas (demand management, load balancing, and event response), corridor operators can develop control strategies (tactics or actions) and procedures for implementing those strategies. For the capital improvement area, corridor operators do not develop control strategies; instead, recommendations for capital expenditures for facility improvements are identified and pursued. Corridor operators may be able to implement some recommendations (e.g., the installation of ITS) more easily than others.

One example of an envisioned strategy for ICM is having the ability to quickly enable load balancing during an unplanned major congestion event on a freeway. For example, if a chemical spill occurs on the freeway, operators would have the capability to quickly enact not only preapproved route shifts to divert traffic onto arterials, commonly known as pre-determined detours, but also to enact preapproved mode shifts amongst the stakeholder groups to initiate remediation without delay. Further, in addition to enacting (and enabling the detours with options and signal progression, et al) the road detours, the stakeholders simultaneously notify travelers of available transit options and encouraging alternate travel times to accommodate their planned trips. It is important to differentiate that this is not traditional "detouring," but rather, a holistic approach to enable the corridor, and not just a route, to absorb the impact of an atypical event, e.g., an hours-long shutdown or newsworthy major event, and not just a local fender bender. This concept makes use of the term "travel shed." Think of a "watershed." In a travel shed all trips (instead of tributaries) combine to absorb the closed freeway and all trips are candidates to be messaged to avoid it in the first place, or pro-actively enabled to get around it. A travel shed, therefore, is defined to be the area subsumed by a cordon, such that all trips therein would otherwise orient to the subject anchor highway if not for the congestion or the propensity for even a minor event on that facility to induce bumper to bumper gridlock, thus requiring the adjacent arterials, cross streets, signals, and systems to mitigate and absorb the subject ICM event. The focus should be on moving travelers through the corridor, not just vehicles. Thus, the real-time advantage of ICM is not simply to identify a detour route, but to inform and empower all corridor travelers to take advantage of the implemented management strategies. The management strategies may include encouraging mode shift, traveler information messages, signal progression, "opening" HOV lanes if necessary, adjusting (or suspending) ramp metering or part-time shoulder use, and/or relaxing peak hour travel and parking restrictions during the duration of the event. Once the event has subsided, and the anchor highway returns to nominal condition, then individual traffic and transit systems return to their normal day to day operation. Travelers are kept informed about the best trip options every step of the way. These types of strategies were demonstrated with varying degrees of success by the Dallas and San Diego demonstration sites.

One final idea; think of ICM in terms of an unplanned special event. A planned special event, like a college or pro game day, or a route-closing parade, requires multi-agency coordination, many stakeholders, and a great deal of pre-planning to mitigate the traffic impact. It effectively shuts down a region, or at least consumes it, sometimes for hours. Consideration is given to changing signals, or opening or closing ramps, or appropriating reverse lanes, or invoking dynamic messaging, and enabling all manner of special activities to absorb the special event, including to promote use of transit (e.g., bus bridges) and subway (where available) to mitigate the crowds. But what if that planned event were to occur, in essence, without planning? ICM, which is most effective in peak periods, and for already-congested corridors, constantly "searches" for atypical congestion, often caused by an immediate non-recurring event or even beyond-normal recurring patterns. The decision support system (DSS — see Appendix A for a fuller explanation) subsequently — and in real-time — recommends and then invokes the best of several alternate mitigation plans, sort of if the "unplanned" event had been planned all along. The many agencies' pre-discussed "business rules" (also explained further below) and mitigations are begun immediately, and not, as in the past, subject to delay from discussion, inaction, and approvals needed up the chain, which would only serve to exacerbate an already-impacting event. The incident still occurs, and is still impacting, but whether a "planned" event, or an "unplanned" incident, but mitigated by ICM, it is respectively lessened.

Pioneer Corridor Sites Development and Demonstration

This section describes the results of the Pioneer site process, including the initial concept development work; the analysis, modeling and simulation (AMS) activity; deployment of their initial concepts; and the independent evaluation.

Confirming the Initial Concept and Baseline Requirements

The Pioneer Sites initially included eight regions (Dallas, Houston, Minneapolis, Montgomery County MD, Oakland, San Antonio, San Diego, and Seattle) that helped to mold the original operational concept of ICM. These regions submitted an initial ConOps and System Requirements outlining their vision of ICM and an ICMS. This initial work helped researchers to develop consensus on what ICM may include. These efforts identified that a successful ICM would include, among other things, the ability to:

  • Manage the corridor as a system rather than individual assets
  • Enable travelers to make informed travel decisions and dynamically shift modes during a trip
  • Reduce travel delays, fuel consumption, emissions, and incidents
  • Improve travel time reliability and predictability
  • Optimize existing transportation infrastructure along a corridor, making transportation investments go farther

Figure 2 depicts the original ICM Pioneer Sites summary of strategies anticipated for their ICMS deployment. The Pioneer Sites ConOps and Requirements work helped confirm what a successful ICMS may include for each region.

Geographic scope and analysis capabilities of AMS tools. Anticipated ICM deployment strategies for each concept development corridor location.  Chart based on information found in surrounding text.
[Key: ● Planning to implement in ICMS]
Figure 2. Chart. Anticipated ICM deployment strategies for each concept development corridor location.
(Source: USDOT)

Highlights of the Analysis, Modeling, and Simulation (AMS) Activity

The first activity for the pioneer sites was defining the initial concepts for ICM strategies in their networks. After this step, three pioneer sites (Dallas, TX, San Diego, CA; and Minneapolis, MN) undertook analysis, modeling, and simulation (AMS) to explore whether applying ICM strategies (such as ramp metering, congestion pricing, signal optimization, transit priority, and enhanced traveler information) to a transportation corridor in a truly active and integrated manner could improve mobility, reliability, and environmental impacts of transportation corridors. The three sites examined the implications of implementing a host of ICM strategies applied under conditions of varying demand along a transportation corridor. The analyses encompassed freeway, arterial, and transit facilities along the defined corridors and examined effects of ICM strategies applied under conditions of high, medium, and low demand. The AMS assessed the effects of ICM strategies both with and without traffic incidents (the largest cause of unexpected congestion) and other scenarios.

Findings across all three sites indicated that ICM would increase reliability and reduce travel time, delays, fuel consumption, and emissions in transportation corridors. Findings also indicated that benefits result when otherwise "stovepipe" agencies combine to cross-share information, resources, and solutions that benefit all. Further, the benefits of ICM appear to scale with travel demand and are especially meaningful under scenarios that unexpectedly constrain supply, such as traffic incidents.

One of the defining features of the ICM AMS methodology is that it helps agencies to understand system dynamics at the corridor level. It uses corridor-level performance metrics (e.g., trip-end travel times and systemic congestion metrics) rather than facility-level metrics (e.g., queues and delays on one facility) to evaluate and understand corridor performance. This is accomplished through the combined use of multiple classes of available modeling tools. Three classes of modeling tools — macroscopic, mesoscopic, and microscopic — are considered essential components of the AMS methodology and were used for this analysis. Figure 3 presents a graphical depiction of the geographic scope and interrelationships between these tools.

This figure depicts the geographic scope and analysis capabilities of the suite of AMS tools brought to bear on integrated corridor management systems. At the highest level is the macroscopic level, capable of modeling at a regional scale (regional patterns and mode shift; transit analysis capability). The mesoscopic scale models the network in more detail, and is capable of modeling traveler information, HOT lanes, congestion pricing and regional diversion patterns. The microscopic simulation modeling unless the network in even greater detail, which limits the geographic scope of which can be represented in the model. The microscopic level allows traffic control strategies such as ramp metering and arterial traffic signal control to be modeled. The 3 levels are shown interacting, and zooming in on different portions of the macroscopic and mesoscopic network.
Figure 3. Chart. Geographic Scope and Analysis Capabilities of AMS Tools
(Source: USDOT, September 2009.)

A major accomplishment of the ICM AMS work is that it generated improved analysis tools and methods for corridor analyses. These tools and methods assist system operators with analyses in five areas: mobility, reliability and variability of travel time, emissions, fuel consumption, and benefits and cost comparison.

The ICM technologies assist operators in gathering expanded data sets useful in managing corridor networks. Through the ICM AMS tools, this data is fused to provide managers with insight on conditions across the full travel shed of the corridor. Operators use this data and work together to implement predefined strategies and coordinate operations to manage the multimodal networks more efficiently and can provide truly "actionable" information to travelers such that they can alter trip (start) times, route choices, and mode choices on a sufficient scale to "soften" congestion hotspots, spreading demand more evenly across the network.

Conducting ICM AMS offers the following benefits5:

  • Invest in the right strategies. The analysis offers corridor managers a predictive forecasting capability that most lack today to help them determine which combinations of ICM strategies are likely to be most effective and under which conditions.
  • Invest with confidence. The analysis allows corridor managers to "see around the corner" and discover optimum combinations of strategies as well as conflicts or unintended consequences that would otherwise be unknowable before implementation.
  • Improve the effectiveness/success of implementation. With this analysis, corridor managers can understand in advance what questions to ask about their system and potential combinations of strategies to make any implementation more successful.

The analysis provides a long-term capacity for corridor managers to continually improve implementation of ICM strategies based on experience.

Summary of the ICM Demonstration Sites

This section contains a summary of the two selected Pioneer ICM demonstration sites (Dallas and San Diego), including a description of the transportation corridor characteristics and needs, project partners, a corridor study area map, the proposed ICM strategies, and a discussion of the status of ICM operations in their regions.

Summary of the Dallas ICM project

A 28 mile stretch of the U.S.75 corridor in the Dallas-Fort Worth region (shown below in Figure 4) was selected as the demonstration site in Dallas.6 The Dallas-Fort Worth region was ranked as the 11th most congested region in the U.S.7, with an expected population growth of one million residents every eight years.

Partners for the Dallas ICM project included:

  • Dallas Area Rapid Transit
  • City of Dallas
  • Town of Highland Park
  • North Central Texas Council of Governments
  • North Texas Tollway Auth.
  • City of Plano
  • City of Richardson
  • Texas DOT
  • City of University Park

The U.S.75 Corridor serves: 1) commuter trips into downtown Dallas, via the freeway, bus routes, light-rail line, and arterial streets; 2) a significant number of reverse commuters traveling to commercial and retail developments in the northern cities and neighborhoods; 3) regional traffic during off-peak periods; and 4) interstate traffic into Oklahoma, since the freeway is a continuation of Interstate 45. The corridor also is a major evacuation route and experienced significant volumes during the Hurricane Rita evacuation in 2005. In Dallas, the inability to expand the freeways or arterials as a method to reduce delays caused by bottlenecks and incidents or to improve travel time reliability created a need to explore alternative congestion reduction strategies. Several features of the corridor study area made it an exemplary ICM testbed: an eight-lane freeway with continuous frontage roads, a concurrent-flow, High-Occupancy Vehicle (HOV) lane, light-rail line, transit bus service, park-and-ride lots, major regional arterial streets within approximately two miles of the freeway, toll roads, bike trails, and Intelligent Transportation Systems (ITS). The layout of the transportation network provided opportunities for strategic traffic diversion onto under-utilized frontage roads, arterials, or transit.

The originally envisioned strategies conceived by the Dallas ICM Team included:

  • Decision-Support System
  • Actionable traveler information
    • Interactive Voice Response (IVR) 511
    • Website
    • Email alerts
    • Comparable travel times
  • Rerouting of traffic
    • Coordinated timing and adaptive signal control
  • Mode Shift
    • Parking management
    • Real-time service adjustments
This figure shows the U.S. 75 corridor network, the site of the Dallas integrated corridor management demonstration. Shown between I 635 in the present George Bush Turnpike is an HOV lane, in orange, and the DART line runs very close to U.S. 75 going north from downtown Dallas. The DART blue and green lines are also shown, coming out of downtown into various neighborhoods to the east of U.S. 75. Yellow highlighting centered on the 75 corridor depicts the primary corridor, while the larger travel shed area is marked with a black marker, and is approximately the shape of a piece of pie, extending all the way from the Dallas North Tollway the Rayburn Tollway to Fairview, and then looping south to State Highway 78 before wrapping back towards the Dallas central business district.
Figure 4. Chart. U.S.75 corridor network, Dallas integrated corridor management demonstration site.
(Source: Google© Map Data, March 27, 2017)

The Dallas ICM team proposed an ICMS consisting of three main components including decision support, data management, and user interface functionality. Their system is still largely in place, although these components have changed from that which was originally envisioned. Perhaps the most visible change is that the Dallas ICM "champion" changed from Dallas Area Rapid Transit (DART) to TxDOT, for reasons that the project centered on the highway — U.S. 75 — over the rail line. The data management system, originally called SmartNet was replaced with a next generation product version called EcoTraffic and is still in operation. The data is shared among network operators through the 511 system. Originally, the corridor data was being collected from seven agencies; however, that has now expanded to over 30 agencies. The real-time ICM data is no longer limited to the U.S.75 corridor subnetwork and is being actively used to support regional operational activities. Additionally, Dallas subsequently received a USDOT Advanced Transportation and Congestion Management Technologies Deployment (ATCMTD) Grant in 2017 that will help to improve access to DART ride-sharing services, thereby improving access to DART stations, and helping to further enhance ICM mode shift capabilities.

Whereas, an automated Decision Support System (DSS) is no longer officially in operation, many of the DSS plan elements are still used to respond to events. During the one-year ICM system test period, the network operators quickly learned which plans would be recommended (i.e., "triggered") for implementation by the DSS based on certain traffic conditions. Once they learned what plans would likely be recommended by the DSS, they did not feel it was cost effective to continue making minor incremental changes to the data model and operating the DSS. Finally, there are no dedicated ICM coordinators or staff. The resultant scaled-down ICM system did not require a full-time staff position for operations. The ICMS coordinator role now occurs in an ad hoc fashion at times of greatest need. The current system changed from the original vision for the ICMS; however, the changes have resulted in a system that is easier and more cost effective to support within the corridor. The Dallas Team has proved that even a scaled down version of an ICM DSS, when used by capable network operators, can effectively manage the corridor.

Dallas stakeholders consider the improved collaboration among network partners to be one of the most valuable outcomes of the Dallas ICM project. The partners gained a greater understanding of each other's operational needs and they learned new operational strategies to better relieve congestion in the corridor. The capture and sharing of the operational data opened the eyes of the network operators to new approaches for congestion relief. Even though the Dallas stakeholders did not maintain or enhance the system as originally intended, the network operators believe that the ICMS is an overall benefit to their corridor. Some of the operators report that they would still like to have the DSS in operation along with a dedicated ICM coordinator. However, funding for a full-time staff position is currently not available.

Summary of the San Diego ICM project

Partners for the San Diego ICM project included:

  • San Diego Association of Governments (SANDAG)
  • California Department of Transportation(CALTRANS)
  • City of Escondido
  • Metropolitan Transit System
  • North County Transit District
  • City of Poway
  • City of San Diego

The I-15 corridor is an eight- to 10-lane freeway, providing an important multimodal connection between San Diego, CA, and destinations to the northeast. It is one of three primary north-south transportation corridors in San Diego County and is the primary north-south highway in inland San Diego County, serving local, regional, and interregional travel. The corridor is a heavily utilized regional commuter route, connecting communities with major regional employment centers. It is located within a major interregional goods movement corridor, connecting Mexico, counties in California, and Las Vegas, Nevada.

The corridor study area, shown in Figure 5, consists of the freeway, including managed/ High-Occupancy Toll (HOT) lanes and general-purpose lanes, frontage roads, bus rapid transit (BRT), park-and-ride lots, and regional arterial streets. The current operations on I-15 include two center-median lanes that run along 8 miles of I-15 between SR 163 in the south and Ted William Pkwy (SR 56) in the north. These center-median lanes are reversible HOV lanes that operate in the southbound direction in the AM peak period and in the northbound direction during the PM peak period. The current operations also allow Single Occupancy Vehicles (SOV) to utilize the roadway for a price, thereby operating as HOT lanes.

Current weekday traffic volumes range from 170,000 to 290,000 vehicles on the general-purpose lanes of I-15; approximately 20,000 vehicles use the I-15 Express Lanes during weekdays. Analysis of corridor conditions show that typical weekday demand along this linear corridor is high, largely due to the limited number of freeway alternatives. Analysis of historical data on this corridor shows that 10 percent of the days in the year experience major incidents under conditions of high demand.

This figure shows the San Diego ICM demonstration corridor, which runs along I 15 from just north of State Route 52 all the way to state Route 78, through Mira Mesa College, Sabre Springs, Rancho Bernardo, Del Lago, and Escondido. The corridor is broken into 3 segments; North segment, middle segment, and South segment.
Figure 5. Chart. I-15 ICM Corridor, San Diego, USA
(Source: USDOT)

The goals of the I-15 ICM initiative were to 1) increase corridor throughput, 2) improve travel time reliability, 3) improve incident management, and 4) enable intermodal travel decisions. Stakeholder agencies defined performance measures to support analysis in areas of mobility, travel time reliability, and emissions and fuel consumption.

The originally envisioned strategies conceived by the San Diego ICM Team included:

  • Automated DSS
  • Actionable traveler information
    • 511 (phone and website)
    • Comparable travel times
  • Managed lanes
  • Rerouting of traffic
    • Coordinated timing and responsive signal operations
    • Coordinated ramp metering and traffic signals
    • Wayfinding roadway signs for diversion routes
  • Mode shift
    • Bus rapid transit
    • Transit signal priority
    • Real-time transit information

The San Diego region implemented and still operates an advanced form of a DSS that enabled automated response plan implementation including ramp-metering controls and signal system timing changes on arterials. The San Diego stakeholders developed business rules and parameters that govern the development of response plans. A response plan generation and evaluation process is triggered by the DSS business rules performance thresholds which can be adjusted and considers loss in capacity (when captured during event and entered by TMC operators via the Caltrans event system) and observed drops in speeds. The system is also designed to start a response plan process manually. The San Diego system predicts the impact of the proposed response plans with the aid of traffic simulation tools and selects the best response plans that meets the system activation and implementation thresholds.

During the ICM demonstration phase, the San Diego region continued to make significant investments in transit, highway, and arterial systems along this corridor to maximize ITS benefits while focusing on data sharing. These investments required the ICM team to continually re-calibrate the decision support system and responses. In some ways, it was difficult to separate the impacts of traditional upgrades in the corridor from "strictly ICM centric" successes. Nevertheless, the San Diego ICM team felt that there were both measurable benefits and institutional benefits to deploying the ICM strategies. The San Diego ICM stakeholders identified needs to optimize operational coordination of multiple transportation networks and cross-network connections to improve corridor mobility within the region. Because the frequency of traffic incidents increases during periods of high demand, the impacts of these incidents are more widespread (i.e., more travelers affected, increased environmental impacts associated with more travelers idling).

The San Diego ICM Team found that the benefit/cost ratio (BCR) for their system goes up as activations go up; however, the ultimate point of corridor operations is to keep the activations low by keeping traffic running smoothly. The ICMS provides network operators with new tools to isolate corridor trouble spots and resolve any congestion related issues in those locations to minimize incident occurrences. This in turn keeps activations lower. Ultimately, the San Diego team has found that ICM is a valuable mitigation tool, and its greatest value may be the improved ability to manage low probability — high impact events.

San Diego also found that interagency collaboration was enhanced and is key to improving corridor operation; this type of collaboration did not exist before the ICM project. For San Diego, ICM is a fundamental change in the traffic management paradigm that has offered a more detailed understanding of the corridor and its operations.

Key Findings from the Independent Evaluation of the two Demonstration Projects

The Dallas U.S.75 and San Diego I—15 projects were the first of their kind in the country and required original research and development prior to implementation. These sites have different characteristics and used different analysis models and methods. Several other regional improvement projects including infrastructure capacity expansion projects were implemented in parallel with the ICM projects in both regions. Thus, it is to be expected that there were broader impacts and improvements that changed the dynamics of the corridor operations before and after ICM. Initially during the evaluation period, only a small number of incidents required ICM response activations. The lack of sufficient activations did not allow for a completely objective empirical assessment of the impacts of ICM. Because of this, the findings from the evaluation were supplemented with the AMS modeling results and other sources of information to meet the goals of the evaluation project. The AMS modeling tools used in San Diego and Dallas were different from each other; specifically, each was promoted and tested on its own merits. The Dallas ICMS used the DIRECT mesoscopic modeling tool while the San Diego ICMS used the (Advanced Interactive Microscopic Simulator for Urban and Non-Urban Networks) Aimsun microscopic modeling tool. The AMS team used these respective tools for conducting their analysis at the two locations, and those results were in turn used by the independent evaluator in the mobility and benefit cost analyses. This makes it difficult, if not incorrect, to draw a direct comparison between the two sites. Readers are encouraged to study the deployments and assess the outcomes of both sites if considering ICMS implementation. The key findings are provided below8:

  • The interagency cooperation and coordination was a big success. Both San Diego and Dallas created a fundamental paradigm shift in the management of their respective corridors by creating strong multi-jurisdictional partnerships that set the foundation for a regional corridor management mindset — based on a platform of strong institutional, technical, and operational integration. The San Diego ICM lead agency, the San Diego Association of Governments (SANDAG) continues to engage regional stakeholders to review and continue to improve ICM performance. Dallas continues to enhance coverage and improve their regional data exchange for information sharing. It is interesting to note that San Diego, led by SANDAG, took a more ambitious approach and was willing to try new things. Meanwhile, Dallas, initially led by DART, and later by TxDOT, took a more conservative approach to using new concepts for real-time traffic management. In the end, both approaches were valid and both teams provided great value to the evolution of ICM.
  • Operators reported better situational awareness of corridor operating conditions, although there were opportunities to improve. Overall awareness of corridor congestion and incidents improved significantly through regional data sharing. Incident reporting improved substantially in both regions compared to the pre-ICM period. However, in Dallas, stakeholders raised the point that there were gaps in arterial data due to outdated equipment and systems issues and that the corridor speed data did not always match with actual field conditions. Operators at both sites believed that construction and maintenance information could have been shared in a more timely and consistent manner.
  • Incident and congestion specific traveler information provision improved. Both sites saw a significant increase in the number of posted dynamic message signs (DMS) and travel time messages post-ICM along with improved incident-related notification to travelers.
  • The DSS at both sites proved to be valuable for better situational awareness, decision-making, and response. A DSS offers impartial evaluation of congestion events and recommends an objective course of action(s) from which to choose. A DSS determines appropriate strategies and responses based on sophisticated performance monitoring and key performance indicators. The DSS detects anomalies as well as return to normal conditions. Additionally, the DSS monitors the availability of corridor devices, provides stakeholders with optimized response plan recommendations, and evaluates the impact of proposed response plans. The DSS at both sites facilitated better awareness of transportation system conditions through the respective data fusion systems and response plans contributing to the provision of better traveler information. The San Diego DSS offered automated response plan implementation across jurisdictional boundaries including ramp-metering controls and signal system timing changes on arterials. The San Diego system predicted the impact of the proposed response plans with the aid of a traffic simulation tool and selected the best response plans that met the system activation and implementation thresholds. The Dallas system initially offered similar features as San Diego, with the exception of their response plan implementation needing to be confirmed by the ICMS coordinator and local agency representatives. Stakeholder agencies in Dallas maintained the authority to decline the requested response plan actions (e.g., signal timing changes) from the DSS. This happened only rarely, but to be fair, one may never know if declining one or more DSS-recommended response plan affected the duration and intensity of an individual event. After one year of operation, the Dallas stakeholders felt that they had learned enough about which plans may be recommended under certain traffic conditions and they chose to discontinue the operation of the DSS, in part due to the cost of maintaining the DSS data model. Dallas learned that manual decision support is a feasible option, if funding for the maintenance of a data model is not available. It should be noted that Dallas did originally intend to ultimately use automated response plan implementation once operators gained trust and a level of comfort with DSS response plan recommendations. The real-time traffic modeling caused some delay in getting response plans implemented. Dallas used a mesoscopic model in part because they had a large travel shed which included a large arterial network. The level of resolution of the network and vehicles is important because the larger the network, the greater the model execution time.
  • The traveler response surveys taken well after the projects concluded indicated mostly positive results on traveler information awareness, utilization, behavioral response, and overall satisfaction. Travelers in both San Diego and Dallas reported higher awareness of where to find traveler information and higher utilization of traveler information post-ICM. San Diego travelers did not report a higher propensity to change behavior in response to traveler information, while Dallas travelers did. While there were some exceptions, generally travelers expressed overall satisfaction with traveler information sources post-ICM. It is empirically difficult for most people to realize and appreciate that what might have been a significant delay (e.g. one hour's delay due to a hugely impacting crash, etc.) was "only" a lesser delay (say, half that) due to the success of an ICM-mitigated event. That person, stuck in traffic for half-again as long as they would normally have been, might not understand that their plight could have been even worse; i.e., they are still "upset at traffic."
  • Alternate route diversion was demonstrated, but transit mode shift did not happen as expected. Overall, alternate vehicle route diversion was a success but transit mode shift (shift from other modes to transit) in the wake of incidents was not observed. The lack of transit mode shift was influenced by inherent constraints in mobilizing (particularly) rail cars and headways, and by not having an immediately available surplus of operators and manpower. Also, regional policies (at both sites) restricted the ability of transit partners to be overly flexible and immediately reactive to ad hoc events, compared to, say, the capability of a highway department or public works department to make on-the-fly tweaks to detours, signals, and diversions. Even beyond these supply-side considerations, the inherent nature of traveler behavior may have prevented meaningful transit mode shifts, particularly when those who normally drive are not familiar with the transit alternative, are concerned about parking, or do not want to leave their vehicles away from home.
  • Corridor mobility performance improved during ICM activations. In San Diego, travel time improved with peak period savings ranging from 250 to 1,300 person-hours during ICM activations. San Diego travel time reliability also improved by an average of 368 hours per Southbound AM peak period activation and 569 hours per Northbound PM peak period activation. In Dallas, travel time marginally improved with peak period person-hour savings ranging from 6 hours to 262 hours. Travel time reliability marginally improved by an average of 109 hours in Dallas for each Northbound PM peak period ICMS activation (there was no travel time reliability improvement in the southbound direction).

    Capturing and validating actual mobility performance improvements proved to be quite challenging for the evaluators. The mobility analysis was driven primarily from the results of the post-deployment modeling and simulation activity, since a before-after analysis using field data was abandoned due to lack of sufficient ICMS activations and the lack of comparable incidents representing before and after ICM deployment. Results also depend on the existing level of saturation within a corridor. Keep in mind that the results in these two sites do not lend themselves to an apples-to-apples comparison. Two different models were used to derive these results and the calibration and validation of the models can make a big impact on the outcomes. Readers are encouraged to look at both methodologies and models carefully before drawing any conclusions.

  • Safety and Air Quality impacts were neutral. Consistent with the primary objective of ICM to improve mobility, the safety and air quality impacts of both demonstration projects were neutral.
  • ICM projects present challenges to traditional benefit-cost analyses and the proper interpretation of results. Using the assumptions documented by the evaluators, the benefits estimated for San Diego's ICMS easily exceeded costs, whereas the break-even point (1:1 ratio) for Dallas's ICMS was contained within the expected range for the benefit cost ratio. For San Diego, the benefit cost ratio ranged from 2:1 to 9:1 (based on the original 17 response plan implementations during the demonstration period) for different scenarios representing different levels of ICMS activation and system effectiveness for a 20-year horizon. For Dallas, the benefit cost ratio ranged from 0.55:1 to 1.64:1 (based on the original 35 response plan implementations during the demonstration period).

    The results for either site could have changed dramatically based on the impacts of the incidents recorded in the corridor. One major incident could have altered the result significantly. ICM also has significant intangible benefits including the successful regional partnerships and the establishment of a joint corridor management mindset, enhanced regional traveler information and 511, ongoing collaboration and information sharing, and traffic signal coordination programs for arterials. These intangible benefits are difficult to quantify and were not considered in the benefit cost analysis. Just as for the mobility analysis, the benefit-cost study depended heavily on the results of the post-deployment modeling and simulation activity. Again, due to the use of different modeling assumptions between the sites, readers should be aware that the benefit cost ratios should not be compared and are not truly indicative of the cost effectiveness of the projects. More information on ICMS benefit cost analyses challenges can be found below in the section on "ICMS Evaluation Benefits and Costs Analysis Insights".

In the end, Dallas and San Diego demonstration sites had different overall outcomes. The original vision for an ICMS was a fully automated software and hardware system that could make automated decisions about real-time operational changes needed to manage congestion within a defined corridor network. These decisions would be guided by real-time network data collected and analyzed by data models incorporated into DSSs for corridor network management.

The San Diego demonstration site achieved a mostly automated DSS that, based on established business rules set by the stakeholders, can generate, evaluate, recommend, and implement a finite set of response plans, with minimal manual intervention. This is in line with the original vision for ICMS implementation and the system seems to work well, although the system required a lengthy period of tuning to find the correct performance improvement thresholds for system activation. In that sense, the human intervention comes in the form of establishing the response plan parameters and decision rules and policies for activating the system.

The Dallas ICMS demonstration varied from the original vision, in that it required human confirmation to implement response strategies that were recommended by the DSS. The Dallas DSS was partially automated and like San Diego's DSS included the ability to analyze, evaluate, and recommend a finite set of preapproved response plans. The Dallas DSS differed from the San Diego DSS by requiring individual operator approval and implementation of recommended response plans. Dallas currently uses human decision support. Dallas demonstrated that the human-in-the-loop solution is a viable option. If ICMS is implemented in this fashion, operations staff need to be instructed, trained, and supported in having ICM responsibilities as a key part of their job descriptions.

The key takeaway from the DSS experiences of San Diego and Dallas is that there is no one-size-fits-all DSS. Each DSS is unique to its corridor and user needs, and needs to account for the advantages and disadvantages of automation vs. human involvement in the choice of the DSS operating model. Many of the ICM deployment planning grant sites are initially focusing on human-centric methods of deploying ICM.

ICMS Evaluation Benefits and Costs Analysis Insights

Review of the demonstration sites independent evaluation revealed several things about ICMS as evidenced from the benefits and costs analysis. Computed benefits depended on DSS activations, and no benefits were derived if the DSS was not activated. Such an assumption under-estimates the benefits of ICMS, including the benefits of decisions that Traffic Management Center (TMC) operators may make outside of the system using their improved situational awareness, which may prevent an unnecessary activation, as well as the benefits that can be attributed to the decisions of travelers responding to the combined, integrated information.

Review of the costs side of the equation reveals the difficulty of properly attributing the costs associated with upgrading network systems to be ICMS-compatible. Costs associated with implementation and operations of the DSS are clearly ICMS-related, as well as any communications upgrades that are necessary to share operational information between infrastructure and service operators and owners. Costs expended to upgrade individual agency networks, however, must be carefully evaluated to determine if the expense should be attributed to ICMS. For example, an interface upgrade that is necessary to connect with the ICMS would likely be counted as an ICMS expense, while the addition of traffic monitoring equipment needed to fill gaps in coverage would likely not be considered an ICMS expense, or at least not exclusively an ICMS expense. While the additional traffic monitoring equipment certainly supports the ICMS operations, its primary purpose is to provide monitoring for the individual network. Analysts must be careful to document these assumptions when conducting a benefits or benefit-cost analysis for a potential ICMS, since the natural tendency may be to underestimate the benefits and overestimate the costs.

Promoting ICM Planning and Deployment

To promote ICM as an operational framework that is routinely considered in cities across the country, the ICM concept needs to be mainstreamed into the transportation planning and programming processes of States, Metropolitan Planning Organizations (MPOs), transit agencies, cities, and other operating agencies. The third major ICM program phase involved a competitive grant and solicitation process9 designed to jumpstart the concept. USDOT sought applications for federal funding (with local match) from multiple sites across the U.S. with candidate corridors suitable for conducting ICMS deployment planning. The solicitation only covered planning activities, with the expectation that funding of future deployment of an ICMS in their regions would be covered by the States and regional/local agencies.

The thirteen sites were selected from 33 candidate proposals. Originally, only ten awards were intended by the announcement, but a congruence of money "asked for" and funds available allowed for thirteen awards. These grant sites were required to produce one or more of the following deliverables as a part of their agreement:

  • ICM Concept of Operations
  • ICM System Requirements Specifications (SyRS)
  • ICM Analysis, Modeling, and Simulation (AMS) Plan
  • ICM AMS Activity Findings Report
  • ICM Implementation Plan

It was intended that once they developed the required pre-implementation documents, they would be ready to move on their own merits and funding towards implementation of their projects. The sites were asked to provide a grant-ending report summarizing the effectiveness of the KTT products that they had access to in supporting their ICM development efforts, as well as provide lessons learned and recommendations for further knowledge transfer activities. Figure 6 on the following page maps the 13 selected sites, shown as red circles, along with the original Pioneer sites. The Pioneer sites taking part in each of the three program stages are shown, including the eight concept development sites (green), the three AMS sites (orange), and the two demonstration sites (blue). Note that the corridors in Minneapolis, Dallas, and San Diego are shown as overlapping circles since they took part in multiple program stages.

This figure shows the United States map location of the Pioneer ICM sites and the planning grant sites. The Pioneer sites taking part in each of the three program stages are shown; including the 8 concept development sites (green), the three AMS sites (orange), and the two demonstration sites (blue). Note that the corridors in Minneapolis, Dallas, and San Diego are shown as overlapping circles since they took part in multiple program stages
Figure 6. Chart. ICM Pioneer Site and Implementation Planning Grant Site corridor locations.
(Source: USDOT, 2018)

After most sites had completed the required deliverable products identified within their cooperative agreements, the USDOT undertook a KTT survey to gather lessons learned and recommendations from the grant recipients. The survey revealed that ICM Deployment Planning Grants were successful in encouraging agencies to systematically plan and implement ICM strategies in the grantee's respective corridors. Grantee sites and stakeholder agencies appear to be committed to some form of ICM deployment; however, there are challenges:

  • Planning and implementing ICM takes time, particularly if an educational process is needed to get everybody on the same page.
  • Funding remains a challenge. Most sites are just getting preliminary implementation funding identified at this point.
  • ICM projects face stiff competition for funding, and must compete against other traditional or transportation systems management and operations (TSMO) projects.
  • Need to determine methods for comparing and measuring traveler-focused, multi-modal performance information.
  • Stakeholders may not understand how to define their "corridor" boundaries, and corresponding travel sheds, to achieve a comprehensive, multimodal, cross-network "system of systems". Some stakeholders will not understand the importance of the travel shed as "the boundary of all last-mile trips that have a high feasibility of using the subject facility (e.g., a freeway)". Experience also reveals that ICM corridors tend to expand in geographic scope as new mode/route alternatives are identified for the travel shed, or traffic patterns change due to the addition of new capacity.
  • Due mostly to funding concerns, it appears that incremental ICM deployment is the most likely viable path forward (rather than a single, large project).
  • Incremental deployment may be preferable from a risk management perspective, but runs the risk of losing momentum as time goes on and personnel change.
  • Related to this uncertainty, many grant recipients were not sure what the end-state ICMS would look like for their corridor.

Ongoing ICM Knowledge Transfer

While previous KTT products and activities (such as guidance, workshops, studies and presentations, articles, and peer exchanges) have been part of every program phase and successful in communicating the ICM concept and promoting deployment, more knowledge transfer activities and resources are needed to mainstream ICM and encourage further adoption. The final ICM program phase places an emphasis on knowledge and technology transfer. KTT products and activities that have been completed to date includes guidance, workshops, studies, presentations, articles, and peer exchanges. Future USDOT ICM work will primarily consist of these kinds of knowledge transfer activities at future new candidate regions, along with focused research for specific questions affecting implementation.

As part of the ICM Program, the USDOT engaged leaders from peer locations implementing ICM today, including representatives from the demonstration sites and early adopter locations, in the development of the content and format for KTT resources to ensure usefulness and practicality. More information on available ICM KTT resources can be found on the ICM Web page at https://www.its.dot.gov/research_archives/icms/index.htm or a related fact sheet at https://www.its.dot.gov/factsheets/pdf/ICM_KTT_V5.pdf.

Resource Type - the following types of KTT resources are available in the ICM Knowledgebase and are linked by topic area below.

Mainstreaming ICM. As mentioned earlier, be advised that we intend to publish an updated companion executive summary called "Mainstreaming ICM: An Executive Level Primer" in late 2019 which intends to address the questions "how do we get started," and "how and why should our region invest in ICM," and, for that matter, "what constitutes a 'candidate' ICM region?" We strongly encourage startup regions to pair this document, which explains to executives what ICM is, with that one that is planned, which will explain how to adopt ICM in your region.

The grant recipients provided the USDOT with valuable information on ICM lessons learned, informational needs, and KTT suggestions or recommendations.

2 https://www.its.dot.gov/research_archives/icms/icm_plan.htm [ Return to 2 ]

3 ICM Knowledgebase is available at https://www.its.dot.gov/research_archives/icms/knowledgebase.htm. [ Return to 3 ]

4 https://rosap.ntl.bts.gov/view/dot/3375 [ Return to 4 ]

5 Source: Integrated Corridor Management Modeling Results Report: Dallas, Minneapolis, and San Diego. [ Return to 5 ]

6 Source: Concept of Operations for the U.S.-75 Integrated Corridor, Dallas, Texas, March 2008. [ Return to 6 ]

7 Source: Schrank, D., B. Eisele, T. Lomax, and J. Bak, 2015 Urban Mobility Scorecard, 2015. [ Return to 7 ]

8 Battelle, Integrated Corridor Management Initiative: Demonstration Phase Evaluation, Final Report Draft, August 2017. [ Return to 8 ]

9 Integrated Corridor Management Deployment Planning Grants, A Notice by the Federal Highway Administration, Federal Register, 2013-26057, 11/01/2013. [ Return to 9 ]

Office of Operations