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

Next Generation of Integrated Corridor Management through Use of Traffic Incident Predictive Analytics and Decision Support System

Project Description

Introduction and Executive Summary

The Tennessee Department of Transportation (TDOT) is a progressive agency that has long been at the forefront of technology adoption. For example, TDOT was the first State to develop and deploy advanced Intelligent Transportation System (ITS) components such as adaptable speed limits based upon automatic detection of severe fog. This culture of advancing technologies, programs, and initiatives is pervasive throughout TDOT and is demonstrated by TDOT's commitment to re-imagining and laying the groundwork for transportation systems of the future – a commitment in keeping with the innovation and technology integration that has been the hallmark of Tennessee throughout the last 200 years. The State's innovative spirit has transformed Nashville from a wilderness outpost to an established innovation hub of the Sun Belt with a gross metropolitan product exceeding $100 billion. This innovation has also resulted in TDOT's recognition as "ground zero" for the next wave of transportation technology and innovation – recognition that is fueled and supported by a shared vision among partners across all levels of state and local government as well as world class academia and research partners such as the University of Tennessee system, Vanderbilt University, the University of Memphis, and Oak Ridge National Laboratory (ORNL).

Tennessee has been a leader in Traffic Incident Management (TIM) as evidenced by the following key actions:

  • TDOT is one of the few state transportation agencies to have reorganized to support Transportation Management & Operations by forming the Traffic Operations Division in 2013.
  • In October 30, 2014, the Tennessee Department of Safety and Homeland Security (DOHS) and the Tennessee Department of Transportation (TDOT) celebrated the opening of the Tennessee Traffic Incident Management (TIM) Training Facility. This one-of-a kind facility features a section of interstate-like roadway ranging from two to six lanes with guardrail, a two-way interchange, and cable and steel barrier rail, as well as a section of two-lane highway and a full four-way intersection. The design provides an area to simulate a variety of crashes, allowing emergency responders to train on safe and efficient techniques for clearing major highway incidents. This training facility has resulted in improvements in the clearance of incidents from roadways in Tennessee.
  • The Tennessee Highway Patrol (THP) has also actively sought more proactive approaches to addressing crashes. In 2014, THP deployed the Crash Reduction Analyzing Statistical History or C.R.A.S.H. system, which utilizes probability analysis to allocate their forces at locations where crashes are determined likely to take place in order to reduce the possibility of occurrence. Troopers are currently using this system for staging of patrol units on Interstates and highways throughout the state to reduce speeds and target aggressive drivers.

The success of the C.R.A.S.H. tool in reducing incidents and improving safety provides an opportunity for an innovative advancement in the way that DOTs manage traffic and incidents. For decades, the transportation community's approach to TIM has been to respond with various Traffic System Management and Operations (TSM&O) or Integrated Corridor Management (ICM) activities once an incident has occurred to relieve the resulting congestion. Because of the advancements in sensing and identifying vehicles as well as "big data analytics" and Tennessee's ability to integrate operations across jurisdictions, we are proposing to change the nature of this relationship between TIM and TSM&O from reactive to proactive. In short, our concept, illustrated in Figure 1, is to utilize a Decision Support System (DSS) using advanced Predictive Analytics (PA) to identify locations and conditions where there is a high probability of an incident and using this prediction to proactively trigger TSM&O actions instead of reacting following an incident's occurrence. This concept enhances our current practices associated with ICM by including the likelihood of incidents as one of the triggers for active management.

Base Foundational Data (Roadside Sensors, Weather, Historical Incident Data) works with System Inputs (SWIFT, C.R.A.S.H., Social Media, Transit Service Available) in the TDOT TMC Central Software System, which leads to System Generated Actions (Advanced Traveler Information Systems, Enhanced Transit Service, Incident Zone Staging and Warning Systems).  This results in the Expected Outcomes: 1) Reduced Traffic Incidents, 2) Reduced injury and fatality of first-responders at Traffic Incident Zones, 3) Mode shift from POV to Transit during times of high-probability of traffic incidents, 4) Reduction of non-reocurring congestion.
Figure 1. The Proposed System Incorporates Predictive Analytics and a Decision Support System based upon Predicting Traffic Incidents as Triggering Events for Traffic System Management and Operations Actions

This project is in clear alignment with the ATCMTD Program and in fact relies upon a wealth of knowledge gained through USDOT sponsored initiatives on Decision Support Systems, Integrated Corridor Management, and TSM&O. This project is only now becoming possible as technology and software systems have recently become able to truly analyze and find complex relationships and associations in data that previously were unidentifiable. This project is proposing to couple advancements in large data analytics and DSS with actionable and impactful activities that will reduce congestion and improve mobility.

Table 1 provides a crosswalk of the proposed technology deployment to the ATCMTD Program expected benefits. As observed in the table, all desired benefits of the ATCMTD Program will be achievable through the proposed project.

Table 1. Project Alignment with Expected Benefits of the ATCMTD Program
ATCMTD Benefits Project Element Delivering Benefit Goal
Reduced traffic-related fatalities and injuries Traffic related fatalities and injuries will be reduced as a result of improved incident clearance times and first responder staging. The safety of first responders working in an Incident Zone will be greatly improved through the proposed Incident Zone System, which includes an early warning and evasion component for first responders.
Reduced traffic congestion and improved travel time reliability Non-recurring congestion due to traffic incidents will be reduced through the Incident Zone System and through mode shift from POV to high-speed, reliable transit. Reliability of transit trips will be improved through the deployment of Transit Signal Priority
Reduced transportation-related emissions As congestion is reduced, so too will emissions.
Optimized multimodal system performance We are proposing as one TSM&O action to initiate a freeway express service that improves travel time reliability.
Improved access to transportation alternatives This project will improve access to transit for commuters from outlying areas into Nashville by doubling the transit availability along I-24. Additional improvements to signalized intersections along Murfreesboro Pike will also improve transit service.
Public access to real time integrated traffic, transit, and multimodal transportation This project includes enhancements to TDOT's existing traveler information and advisory system, Smartway, to include predictions of non-recurring congestions due to incidents and accidents. This information will be made available through a web portal for consumption by the traveling public.
Cost savings to transportation agencies, businesses, and the traveling public Non-recurring congestion due to incidents on I-24 is a significant issue resulting in significant emissions, fuel expenditures, and travel delays by the traveling public. Addressing this source of congestion proactively will provide significant cost savings.

Description of Applicants

TDOT will be the lead fiscal agency for this grant and will ultimately be responsible for overall project delivery. TDOT is partnering with several cities in the study area, particularly those along I-24 and Murfreesboro Pike from Murfreesboro to Nashville. These municipalities will be responsible for managing project elements deployed within their respective jurisdictions. Additional partners include Vanderbilt University and Oak Ridge National Laboratory (ORNL) among others as illustrated in Figure 2 and discussed below.

Organizational Structure.  The US DOT ATCMTD working with the Tennessee Transportation Technology Corridor Program management Office will partner with the Supporting Organizations and Partners (Vanderbilt University, Nashville, City of Murfreesboro, Smyrna, La Vergne, Oak Rick National Laboratory, and the Tennessee State Trooper Highway Patrol).  Implementation will be done for Incident Zone Applications (TDOT, TSHP), ATIS (TDOT), FreeWay Express Transit Service and BRT (MTA/RTA), and TIM Predictive Analytics (Vanderbilt, Oak Ridge).
Figure 2. This Project will be Led by TDOT in Partnership with MTA/RTA, Vanderbilt University, Nashville, Murfreesboro, Smyrna, La Vergne, ORNL, and the Tennessee Highway Patrol

The Tennessee Department of Transportation (TDOT) provides multimodal services for roadways, aviation, waterways, and railroads throughout the State of Tennessee. In addition to planning, operating its own network of over 14,000 miles, TDOT provides funding and technical assistance to cities and counties throughout Tennessee in planning and constructing their roadways. TDOT has long been at the forefront of technology adoption and was the first state to develop and deploy advanced ITS components such as adaptable speed limits based on automatic fog detection. Tennessee has also been a national leader in traffic incident management and is home to the Tennessee Traffic Incident Management Training Facility.

The Nashville Metropolitan Transit Authority (MTA) provides bus, bus rapid transit, paratransit, and park and ride services for the City of Nashville and Davidson County. MTA serves a daily ridership of about 30,000 with a system of 137 busses and 40 paratransit vehicles. Most bus routes serve Music City Central, the downtown transit station, with limited crosstown service available. MTA was among the first transit agencies to implement a fare box that accepted credit cards, though this approach has been discontinued. MTA, in partnership with the Regional Transportation Authority of Middle Tennessee, recently engaged in a region wide public dialogue about the best strategies to improve mobility in anticipation of more than 1 million new residents expected throughout Nashville and Middle Tennessee by 2040.

The Regional Transportation Authority of Middle Tennessee (RTA) provides transit services to the municipalities outside of Nashville and Davidson County, including Franklin, Gallatin, Hendersonville, La Vergne, Smyrna, Murfreesboro, Spring Hill, Joelton, and Springfield. This is accomplished through a combination of express regional bus services contracted to Gray Line of Tennessee and commuter rail service via the Music City Star. This rail service utilizes a 32-mile section of track belonging to the Nashville & Eastern Railroad Authority with three trains providing weekday morning and evening service to six stations.

Vanderbilt University is an internationally-ranked private research university in Nashville with approximately 12,000 students from all 50 states and over 100 foreign countries. The Vanderbilt School of Engineering includes programs in the fields of biomedical, chemical and bimolecular, civil and environmental, construction, cyber-physical systems, electrical and computer science, and mechanical engineering. Vanderbilt operates nearly 70 interdisciplinary research centers including the Institute for Software Integrated Systems, which has developed a wide range of software-intensive systems from small embedded devices, through real-time distributed systems, to globally-deployed complex systems. Additionally, the Vanderbilt Initiative for Smart City Operations Research aims to create a trans-institutional center focused on multi-disciplinary research for formulating, developing and deploying "smart city" applications in Greater Nashville.

The City of Nashville is the capital and largest city in Tennessee. With a 2016 population approaching 700,000 residents, Nashville is at the center of a combined statistical area of 1.95 million people. Historically the home of country music, Nashville has evolved into a major economic magnet and has seen considerable population growth in recent decades. The region is expected to add more than 1 million new residents by 2040. The City of Nashville has demonstrated its commitment to transportation technology though its vision statement for the USDOT Smart City Challenge. Their proposal included smart parking, municipal fleet conversion to electric vehicles, ridesharing, partnerships with the automotive and technology industries to test autonomous vehicles, and open data information technology applications.

The City of Murfreesboro was first established in 1811 and was recognized in 1817 by the State Legislature as a state city. With a population of approximately 131,000 residents, Murfreesboro is the 10th fasting growing cities in the State. Originally rooted in agriculture, the city now has a diversified business base with major employers including Nissan Motor Manufacturing (6,200 jobs), Ingram Book Company (2,300 jobs), and Whirlpool (2,000 jobs). Murfreesboro is home to Middle Tennessee State University and serves as a bedroom community with significant commuter to Nashville from Murfreesboro.

The Town of Smyrna is located between Murfreesboro and La Vergne along Murfreesboro Pike and has approximately 45,000 residents. The Nissan Smyrna Assembly Plant is the largest employer in the 23 square mile town providing roughly 6,500 jobs.

The City of La Vergne is located between Smyrna and Nashville along Murfreesboro Pike and covers approximately 21 square miles. The population in La Vergne is estimated to be approximately 35,000.

Oak Ridge National Laboratory (ORNL) is one of the Department of Energy's premier National Laboratories. Located in Oak Ridge Tennessee near Knoxville, ORNL is home to several of the world's top supercomputers including the world's third most powerful supercomputer. The National Transportation Research Center (NTRC) at ORNL is a comprehensive transportation technology facility and is a DOE-designated User Facility that leverages ORNL's world-leading science capabilities to address national transportation challenges. For this grant, ORNL's expertise in high performance computing coupled with big data analytics and a fundamental understanding of the nuances of the transportation system will be critical.

The Tennessee Highway Patrol (THP), a division of the Tennessee Department of Safety and Homeland Security, is responsible for enforcing federal and state traffic laws on the state-owned roadway network. Agency responsibilities include assisting motorists, investigating traffic accidents, and prosecuting cases involving the use of drugs or alcohol as a contributing factor to an accident. In response to the increasing number of first responders that have been struck and killed while working in traffic incident zones, THP developed the Crash Reduction Analyzing Statistical History (C.R.A.S.H.) system, which has been successful in reducing incidents and improving safety on Tennessee's roadways.

Project Area

Nashville sits at the crossroads of three interstates (I-24, I-40, and I-65), which has enabled considerable population growth both within city limits as well as in ex-urban counties. Regional growth and travel trends and preferences have resulted in an overburdened infrastructure and a decrease in mobility. More than half of all commuters in the Middle Tennessee region work in a different county from which they live and 81.3% of residents commute alone. In 2014, the average auto commuter lost 45 hours due to traffic, costing the region more than $800 million. These problems will worsen with time as the region adds another 1 million residents by the year 2040. Trends such as the ones above are particularly pronounced on the Murfreesboro Pike/I-24 Corridor, which has seen a rapid growth in travel.

The I-24 corridor between Murfreesboro and Nashville carries significant traffic volumes, with the volume-to-capacity ratio often approaching 1.0 during typical peak periods. When incidents that close a lane or the entire roadway occur, travelers often seek alternate routes. On the segment of I-24 between Exit 52 in Davidson County and Exit 78 in Rutherford County, this traffic diverts from I-24 onto Murfreesboro Pike. Unable to handle this volume of traffic, gridlock typically occurs in these conditions.

The cities between these roadways see a high concentration of economic activity, with numerous warehouse and automotive manufacturing facilities including plants operated by Nissan and Bridgestone, both of which are among the largest employers in the region. The I-24 corridor sees significant commodity flows, particularly in the construction and mining and mixed/warehouse shipments commodity groups. The weight of construction and mining shipments is projected to grow by 36% from 82 million tons in 2010 to 112 million tons in 2040, while mixed/warehouse shipments whose tonnage is expected to grow by 192% from 20 million tons to 62 million tons. With a current mode split of 43% truck in 2010 and a projected split of 50% in 2040, truck traffic will continue to grow substantially on the I-24 corridor.

I-24 currently utilizes an HOV system that is approximately 50 miles in length, running from US-231 in Rutherford County to Harding Place in Davidson County, approximately 8 miles south of downtown Nashville.  These lanes are signed and striped but there is no barrier separating them from general purpose lanes, nor do they have separate access or egress. In general the public has expressed favorable comments about these HOV lanes.

MTA currently operates a bus rapid transit (BRT) lite route on Murfreesboro Pike between downtown Nashville and Antioch, with 15-minute headways. Additionally, RTA currently operates three routes from Murfreesboro to Nashville including route 84X, which operates on I-24 using HOV lanes; route 96X, which operates on Murfreesboro Pike; and route 86X which uses a combination of the two. Recent outreach conducted as part of the I-24 Corridor Study indicates that the public favors new and enhanced transit options as one means of improving mobility along the I-24 corridor. The public also favored new and improved park-and-ride lots paired with managed lane technology on I-24 as another potential strategy to reduce congestion.

Vision, Goals, and Objectives of the Project

The vision of TDOT and Partners with this project is to dramatically improve integration of TIM into Integrated Corridor Management and overall TSM&O practices. This will be accomplished by developing and deploying an advanced Predictive Analytics and Decision Support System that will provide predictive incident probabilities with enough fidelity and accuracy so that preventative actions can be taken to reduce or eliminate non reoccurring congestion caused by an incident, as well as improving the safety of first responders working in an incident zone.

In many respects, our vision aligns with the USDOT's vision for the Response, Emergency Staging and Communications, Uniform Management and Evacuation (R.E.S.C.U.M.E.) Dynamic Mobility Applications bundle. For example, as in the R.E.S.C.U.M.E. bundle, we are proposing a first responder warning system for Incident Zone Management along with the dynamic creation of an electronic incident zone using a combination of radar and DSRC equipment. Also included in our vision, however, are other activities such as enhancements to transit to reduce the overall number of vehicles on the highway as well as providing advanced traveler information systems regarding traffic and mode alternatives. Our vision, therefore, is to create a holistic environment where incidents are predicted prior to their occurrence and through staging of first-responders, notifying the traveling public, and offering real transit alternatives the incident is avoided and never actually occurs.

This vision is accompanied by several goals and objectives that if achieved will significantly improve the travel experience of both residents and visitors utilizing I-24 and Murfreesboro Pike. Table 2 summarizes these project goals and links them to objectives and the proposed technology solutions that are described in more detail in the following section.

Table 2. Project Goals and Objectives
Project Goal Objectives Technology Deployment Component
1 2 3 4
Reduce Non Reoccurring Congestion Due to Incidents
  • Improve prediction of incidents so that the predictions are accurate, precise, and reliable
  • Establish triggers for system management activities based upon predicted probabilities of an incident
  • Provide traveling public with information on probability of a major incident through ATIS
No Value
Improve Travel Time Reliability and Availability of Transit
  • Adding capacity to existing transit service utilizing the highway
  • Extend TSP on main arterial (Murfreesboro Pike)
No Value No Value
Improve Safety of First Responders in Incident Zones
  • Implement Automatic Establishment of an Incident Zone through DSRC and Radar
  • Provide Incident Zone Intrusion Warnings to First Responders in Time for Evasive Actions
  1. TIM Decision Support System
  2. Smartway ATIS
  3. Enhanced Transit Service
  4. Incident Zone Staging and Warning System

Table 3 provides a cross-walk of the goals and desired focus areas of USDOT with the planned project. As observed in the table, this project will accomplish many of the goals across several of the desired focus areas.

Table 3. Summary of how MOBL-T will meet the Goals of the ATCMTD Program
Goals of the ATCMTD Program How Project will Meet Goal
Reduced costs and improved return on investments, including through the enhanced use of existing transportation capacity We will leverage existing tools such as the C.R.A.S.H, SWIFT, and roadside sensors for the DSS system.
Delivery of environmental benefits that alleviate congestion and streamline traffic flow The focus of this project is on prevention of non reoccurring congestion due to major incidents.
Measurement and improvement of the operational performance of the applicable transportation networks The focus of this project is on prevention of non reoccurring congestion on I-24 due to major incidents as well as reducing POV through enhancements of the transit system.
Reduction in the number and severity of traffic crashes and an increase in driver, passenger, and pedestrian safety By staging first responders at locations and times of high probabilities of incidents, we seek to reduce the number of incidents.
Collection, dissemination, and use of real time transportation related information to improve mobility, reduce congestion, and provide for more efficient and accessible transportation We are proposing to provide the traveling public, via the Smartway tool or a similar mobile application, another source of information on which to base travel decisions: Predicted Probability of a Major Incident.
Monitoring transportation assets to improve infrastructure management, reduce maintenance costs, prioritize investment decisions, and ensure a state of good repair This goal is not directly addressed by the proposed project.
Delivery of economic benefits by reducing delays, improving system performance and throughput, and providing for the efficient and reliable movement of people, goods, and services TDOT has estimated that travelers along I-24 lose approximately $1,600 annually due to congestion induced delays. This project will enhance mobility and reduce these costs by minimizing the non reoccurring congestion due to major incidents.
Accelerated deployment of vehicle-to-vehicle, vehicle-to-infrastructure, and automated vehicle applications, and autonomous vehicles and other advanced technologies DSRC equipment will be used within the incident zones in combination with radar units to detect and warn of oncoming vehicles that may present a hazard to first responders in the incident zone.
Integration of advanced technologies into transportation system management and operations We are proposing a paradigm shift in the process of TIM and its integration into TSM&O by using the predicted probability of an incident as a triggering action rather than attempting to react following an incident.
Demonstration, quantification, and evaluation of the impact of these advanced technologies, strategies, and applications towards improved safety, efficiency, and sustainable movement of people and goods This project will provide data on the effectiveness of using a DSS system to predict non reoccurring congestion and the feasibility of taking action to prevent or reduce undesirable outcomes from occurring.
Reproducibility of successful systems and services for technology and knowledge transfer to other locations facing similar challenges. If successful, this program could provide a roadmap for similar systems throughout the U.S. Other locations, such as Ohio, are already replicating aspects of the C.R.A.S.H. system and would be good candidates for future adoption. The technologies being proposed will only improve during the next decade with further advances in computer processing and advancement of Connected Vehicle and Autonomous Vehicle technologies.

Technology Solutions and Transportation Services

The proposed project will address safety, mobility, and congestion as described above through the application of several technology solutions. Cornerstone to our proposed project is the development and implementation of a DSS for predicting the probability of congestion causing incidents on I-24. Once these predictions have been generated, they will be used to trigger additional actions, each involving a degree of technology and operational components. The following seven technology-based solutions will serve as the primary components of this project:

  1. Centralized Traffic Management Center Software System
  2. Predictive Analytics and Decision Support System for Traffic Incidents
  3. Advanced Traveler Information Systems (SmartWay)
  4. Enhanced Freeway Transit Service with Increased Capacity
  5. Enhanced Transit on Major Arterial Through Expansion of Transit Signal Prioritization
  6. Pre-Incident First Responder Staging Based Upon Predicted Incident Probabilities
  7. DSRC and Radar Based Incident Zone Safety Systems.

Although Tennessee is a progressive state and has the ability to shift the current state-of-the-art for incident management through this grant from reactive to proactive, Tennessee also has the ability to share lessons learned and to train others on processes and techniques. Our traffic incidents, and the resulting non reoccurring congestion that ensues can be found in all States throughout the country. We believe that the collection of technologies provides the best opportunity to improve safety and mobility, but individually each of the technology solutions can be readily adopted by other States. To improve the transferability, we are leveraging the expertise and tools of ORNL, which are the same tools and resources that would be available to every State. The DSS will not be a proprietary system – it will be open source and made available through the USDOT's Research Program and through ORNL. The following describes in more detail each of the proposed technology components.

Centralized Traffic Management Center Software System.

One of the factors that make TDOT and Tennessee an ideal State for this application is the ongoing development and integration of all Traffic Management Operations throughout the State into a common, statewide Traffic Management Center Software System. TDOT will be implementing a customized version of Southwest Research Institute's (SWRI) ActiveITSTM traffic management software (see Figure 3). This central traffic management software system will provide a unified platform for TDOT to share with local agencies across the state. The singular platform will assist TDOT with creation of Center-to-Center deployments across the state and will allow for sharing of data and resources between TDOT and Local agencies, which will help federal dollars go further. This system includes the capability to monitor and manage literally thousands of ITS components and, more importantly, automates incident management operations through automated event management response plans.

Screenshot of ActiveITS website.
Figure 3. TDOT will be Adopting a Customized Version of SWRI's ActiveITS Platform to Serve as a Centralized Traffic Management Center Software System Across the State. This software system will house the TIM DSS.

The TIM Predictive Analytics Engine and DSS will be housed within a customized module in ACTIVEITSTM, ensuring its availability and applicability across the State. This will enable a straightforward expansion of the deployment to other corridors in both urban and rural areas of the State following the initial development and deployment of the technologies and TSM&O activities on I-24 and Murfreesboro Pike as part of this grant project.

Predictive Analytics and Decision Support System for Traffic Incidents

The Predictive Analytics and Decision Support System for traffic incidents will be a customized module residing within the ActiveITSTM software. The advancement in sensors and information transfer technologies provide opportunities to collect large amounts of data on complex operations and processes that govern various aspects of travel. In such situations, where it is an almost intractable task to build complex models, data driven approaches can produce more informed solutions to problems than using heuristics and ad-hoc approaches.

This project will develop scalable computing, data, and analytics framework to discover alternative methods to manage urban traffic congestion. The transportation system is beginning a Connected and Automated Vehicle revolution, where communication between vehicles and between vehicles and the highway infrastructure are proliferating. The unprecedented abundance and evolution of data, scalable computing, and cyber infrastructure have propelled the science of data analysis and visualization. This presents the opportunity to derive novel insights from the interconnected paradigm of transportation and energy. For knowledge discovery, characterization of the interactions between the human dynamics and energy and transportation infrastructures are essential and requires integration of three distinct components, namely, data, models and scalable computation. In particular, over the last years, deep reinforcement learning has disrupted many applications in robotics, machine vision, finance, and many other fields. With the massive amount of data on energy consumption and mobility patterns, similar breakthroughs are expected to happen for energy efficient mobility with a minimum of traffic congestion.

In this project, we are proposing to develop an innovative PA and DSS that has the defined objectives of determining the expected probabilities of a future major congestion inducing traffic incident on I-24 based upon current conditions and historical trends. Historical information such as previous locations, times, severity of crashes for each segment of I-24 as well as historical traffic volumes, speed profiles, and other information will be used. This historical traffic information will be combined with other non-traditional sources such as social media, cell-phone route requests, weather, and other data through statistical models and deep machine learning techniques to develop a predicted probability of a future incident. Figure 4 illustrates an overall approach to using big data for such types of analyses and predictions.

Approach to using big data.  Inputs (TDOT data, USDOT data, Federal and State Transportation data Sets, Private Sector Data) is used in virtual discovery testbed that uses national scale simulation and machine learning algorithms and this results in reduced traffic congestion, safety of public and first responders, data driven policy making, and an approach to freight movement and personal mobility optimization.
Figure 4. Illustration of the overall approach to using big data, large-scale simulations, and machine learning algorithms to accelerate national mobility and energy infrastructure optimization.
Source: ORNL

Advanced Traveler Information Systems

Screenshot of SmartWay TDOT
Figure 5. TDOT's Current ATIS SmartWay will be Enhanced to Show the Predicted Probabilities of a Future Congestion Inducing Traffic Incident on I-24 Along with Availability of Transit Service

TDOT has previously developed and operates an Advanced Traveler Information System called SmartwayTM (see Figure 5). SmartWay provides up-to-date traffic information on Tennessee's highway system. The system currently includes four transportation management centers, 517 cameras, 174 message signs, 1015 roadway detection systems and 49 video detection systems in the four largest cities. Available to the public through a web-site interface that can also be viewed on mobile devices, the SmartWay provides current information for motorists on highway incidents, construction activities and traffic information. It provides views of TDOT's cameras, DMS board information and speed data.

As part of this project, TDOT will upgrade the SmartWay system to include the future predictions of the probabilities on congestion inducing incidents as another data information layer. Accompanying this will be information on the availability of transit alternatives and an indication of whether the probabilities of a traffic incident have triggered transit signal prioritization or other TSM&O actions. Providing this information along with real-time information on transit availability and location provided to the system by MTA/RTA will assist pre-trip travelers in choosing transit as an alternative to POV during periods where there is a high probability of a congestion inducing incident on I-24.

Enhanced Freeway Transit Service with Increased Capacity

The Middle Tennessee Regional Transportation Authority (RTA) was established in 1988 to provide ridesharing and rapid transit service from outlying areas into Nashville. This service includes service between Murfreesboro and Nashville along both Murfreesboro Pike and I-24 (see Figure 6).

Map of existing transit service between Murfreesboro and Nashville.
Figure 4. Existing Transit Service between Murfreesboro and Downtown Nashville

This service includes three transit routes denoted by 84X, 86x, and 96x. The 96x route follows Murfreesboro Pike and will be operated as a Bus-Rapid Transit (BRT) Lite service from Nashville to Bell Road. The 84X and 86X both utilize I-24 as a freeway express service from Nashville to La Vergne where the 86x route diverges to Murfreesboro Pike. These are popular transit routes with heavy utilization and ridership. However, travel time reliability, particularly on days where there is a major incident on I-24 can be challenging. To continue to promote the use of this express transit service, thereby reducing congestion associated with POV travel, we are proposing to increase the capacity of these existing routes by adding an additional four transit vehicles into service. At the same time, as part of TSM&O actions during times where there is a high probability of a traffic incident, we will incentivize the use of this expanded service through fare discounts and/or waivers.

Enhanced Transit on Major Arterial Through Expansion of Transit Signal Prioritization

Currently, the 96X route is only planned to be operated as a BRT Lite between Nashville and Bell Road. This grant will allow RTA and the grant partners to accelerate the extension of this BRT Lite along Murfreesboro Pike. The current planned upgrade along Murfreesboro Pike includes Transit Signal Priority (TSP) at 41 intersections along the Murfreesboro Pike Corridor from Lafayette Street & Wharf Avenue to Bell Road & Morris Gentry Boulevard. New traffic signal controllers and signal equipment along with new fiber along the corridor. Pedestrian improvements at 39 intersections along the Murfreesboro Corridor from Lafayette Street & Wharf Avenue to Bell Road & Morris Gentry Boulevard are also included. As part of this TIGER grant, 41 transit vehicles will be equipped with TSP equipment. This project will be completed in January 2019.

Pre-Incident First Responder Staging Based Upon Predicted Incident Probabilities

Although the existing C.R.A.S.H tool only provides probabilities of incidents within a 30 square mile area in 4 hour increments it is currently achieving a 72% accuracy rate. The Tennessee State Highway Patrol has begun to utilize C.R.A.S.H. predictions to stage patrol vehicles at locations where the tool indicates that there is a high probability of an incident. The presence of a patrol vehicle and officer on the roadway results in fewer speeding vehicles and helps to prevent distracted driving.

Under this grant, we will be greatly enhancing the C.R.A.S.H. predictions to be specific to segments along I-24 from Murfreesboro to Nashville. As such, this will provide even more specific guidance to the Highway Patrol for staging of officers at locations along I-24 with high probabilities of a future incident. Additionally, TDOT will also stage their fleet of Incident Response Vehicles at or near on-ramps to these same highway segments. This staging should enable TDOT to arrive quickly on-scene to establish an added safety zone for the Highway Patrol officers and other first responders working the incident.

DSRC and Radar Based Incident Zone Safety Systems

view of back of a traffic signal work truck.  Highlighted are the Radar Unit and the Connected Vehicle Radio.
Figure 7. TDOT will Implement the INC-ZONE Component of the R.E.S.C.U.M.E. Dynamic Mobility Application Bundle Enhanced with Vehicle Mounted Radar

One of the key technologies proposed for deployment is radar and DSRC equipment mounted on TDOT Incident Response Vehicles and worn by incident first responders to provide both situational awareness and an early warning for responders of vehicles breaching the incident zone boundaries. In short, we are proposing to implement the Incident Zone component of the R.E.S.C.U.M.E. Dynamic Mobility Application Bundle, but enhance the DSRC technology with vehicle-mounted radar systems (see Figure 7).

Due to the pre-incident staging described above, we expect that TDOT will be able to arrive at a traffic incident quickly and immediately establish an electronic Incident Zone using the Radar and DSRC equipment mounted to the vehicle. Traveler Advisory Messages will be broadcasted via the J2735 Message Set to on-coming vehicles. Using both the Basic Safety Messages (BSMs) of on-coming vehicles and the radar units mounted to TDOT's Incident Response Vehicles, on-coming vehicles will be tracked in real-time as potential threats to on-scene first responders. Incident Zone warnings and alerts will be issued to oncoming vehicles as appropriate. If the oncoming vehicle is determined to be a threat to first responders (e.g., demonstrating erratic or districted driving patterns, incursion into the incident zone, etc.) an alert will be sounded on the incident response vehicle and broadcasted to equipment worn by on-scene first responders. For example, potential mechanisms for providing alerts to first responders would include utilization of a connected trunk radio, a DSRC backpack radio, or a DSRC enabled beacon or vest (see Figure 8).

Images of potential personal alert systems.  Included are safety vest DSRC Backpack, radio, traditional trunked radio.
Figure 8. Potential Personal Alert Systems Include DSRC Enabled Safety Vests, DSRC Backpack Radios, and the Traditional Trunked Radio

Deployment Plan

To accomplish the goals and objectives of the proposed system, we have organized activities and aligned schedules using a detailed Work Breakdown Structure (WBS). A high-level version of the WBS is presented in Figure 9 with details for each task contained in the following sections.

High-level work breakdown.
Figure 9. High-Level Work Breakdown Structure for the Proposed Project

Task 1. Project Management

This project will be led by Mr. Phillip (Brad) Freeze, who will be responsible for all aspects of the project, including preparing a detailed Project Management Plan (PMP) that includes both a more detailed WBS as well as a detailed project schedule. Development of the PMP will be one of the first tasks initiated on this project. Regular status updates will be provided to USDOT annually as required by the grant and periodically as needed and desired by USDOT.

Task 2. Service and Equipment Acquisition

Following the completion of the PMP, TDOT will engage and begin the processes of acquiring the equipment and services needed to implement the project. This includes establishing subcontracts and memorandums of understanding with the project partners, obtaining the DSRC and Radar Equipment, obtaining the TSP equipment needed to extend TSP on Murfreesboro Pike, and the transit vehicles to increase the capacity of the three RTA express bus routes. The following lists the specific equipment and quantities that will be procured:

  • 22 Vehicle mounted Radar and DSRC Units for TDOT's Incident Response Vehicles
  • 40 DSRC personal protection gear for First-Responders working in Incident Zones along I-24.
  • 4 Full-Size Transit vehicles
  • 20 sets of Transit Signal Priority equipment for Signalized Intersections

We realize that some of this equipment, such as the transit vehicles, have a long lead time and we will incorporate these lead times into the overall project schedule. Some of this equipment will be procured by TDOT while others may be procured more readily by one of the project partners.

Task 4. System Development

There are four sub-systems that will be developed or modified under this grant as illustrated in Figure 1. The following summarizes our approach for either developing or modifying each of these four sub-systems.

Centralized Traffic Management Center Software System

TDOT staff, with assistance from SWRI, will be responsible for the installation and development of the centralized traffic management center software system. As indicated before the centralized software will be based upon the ActiveITSTM system first developed by SWRI. This software was selected by TDOT following an extensive research and selection process and will be installed on servers within TDOT Traffic Management Center.

Predictive Analytics and Decision Support System for Traffic Incidents

Development of the Predictive Analytics and Decision Support System based upon Predicated Traffic Incidents will be the responsibility of Vanderbilt University with support from ORNL. In prior work, Vanderbilt professors have described integration methods for spatio-temporal incident forecasting using previously collected vehicular accident data provided to us by the Nashville Fire Department. The literature provides several techniques that focus on analyzing features and predicting accidents for specific situations (specific intersections in a city, or certain segments of a freeway, for example), but these models break down when applied to a large, general area consisting of many road and intersection types and other factors like weather conditions. ORNL expertise in data research to manage data sets for regional traffic simulations in an intelligent mobility corridor will provide a foundation to extend the current-state-of-the-art to a more widespread, regional coverage. This research will further require ORNL artificial intelligence, machine learning, and deep learning algorithms for traffic prediction and congestion management at a regional scale.

Our approach will use a novel data-driven incident-prediction and resource optimization tool-chain that will be developed and deployed in TN as part of this project. Towards this purpose, we will first develop mechanisms to aggregate heterogeneous historical data related to incidents, weather, traffic conditions, road conditions, and transit demand via the corridor. Then, we will analyze them to develop predictive models that will help the TDOT to efficiently allocate and route emergency vehicles, and start on-demand express transit-services, as needed. Thereafter, they can use novel social-media, and mobile application-based information dissemination and notification mechanisms to reduce the likelihood of principal features reaching the threshold criteria, and reducing the likelihood of anticipated incidents before they occur.

A practical solution for building a decision support system is a Bayesian inference tool-chain that can enable the prediction of likelihood of response features as a function of values of the predictor features. For this purpose, we will build upon our prior work of incident prediction tool-chain to learn from other features that are available as part of the TDOT information management system. Towards optimizing resource allocation and reducing the overall incident response time we will develop a mathematical program that will maximize expected incident coverage, and develop a Markov decision process based algorithmic framework for solving the optimization problem. Prior art in incident prediction does not generally consider incident priorities that are crucial in optimal dispatch, and spatial modeling either considers each discretized area independently, or learns a homogeneous model. In our approach, we will bridge these gaps by learning a joint distribution of both incident arrival time and severity, with spatial heterogeneity captured using a hierarchical clustering approach. This approach groups incidents with similar characteristics together, creating a prediction model per group, making forecasting for each group more accurate.

To the extent possible, the PA and DSS will be developed as a module that can integrate into ActiveITSTM. However, due to the extensive nature of the data and the possibility that high-performance computing will be required to enable timely predictions of the probabilities of future incidents, the PA and DSS may be hosted at either Vanderbilt or ORNL with direct ties back to ActiveITSTM.

Advanced Traveler Information Systems

TDOT previously developed the SmartWay system as a RESTful web service based upon the use of a Representational State Transfer (REST) architecture. This allows the service to be accessed across multiple platforms and browsers. Each set of data is presented as a layer of information that a user can select or unselect as they desire. We will enhance this ATIS by adding additional layers that show the current and future predictions of congestion inducing traffic incidents overlaid on I-24 and Murfreesboro Pike. Another data layer will include the real-time location and availability of the RTA Route 84X, 86x, and 96x transit vehicles. Overall information and travel suggestions to travelers, including the potential use of transit as an alternative, will be provided using overlays and popup window boxes. SmartWay will be enhanced by internal TDOT staff.

DSRC and Radar Based Incident Zone Safety Systems
Diagram of a car approaching a safety zone with reduced speed.  The Responder receives personal alert tones if they are approaching, in violation, or are in the warning zone before a an incident.
Figure 10. TDOT will Develop the Inc-Zone using Algorithms and Principles Developed and Demonstrated for the R.E.S.C.U.M.E. Prototype.

Outside of the PA and DSS, this technology component will require the most development and customization by the partners. We will develop this application consistent with the R.E.S.C.U.M.E. Inc-Zone application, leveraging the software previously developed by USDOT. We expect to modify this software to accommodate simplified operational constraints imposed by only mounting the radar and DSRC equipment on TDOT's Incident Response Vehicle in comparison to all first –responder vehicles as implemented in the R.E.S.C.U.M.E. prototype. We are expecting to develop additional customization to account for the addition of a radar vehicle detection to improve safety of the workers during the interim period when connected vehicle technologies are not ubiquitous. As illustrated in Figure 10, our algorithms for determining when a vehicle poses a potential threat will be based upon the R.E.S.C.U.M.E. electronic fencing created by the DSRC equipment, but will be enhanced to include triggers based not only upon vector of approach and speed, but the observed pattern of behavior of the vehicle that would indicate a distracted or drowsy driver (i.e., lane drifting, sharp steering corrections, etc.).

Task 5. Design and Implementation Plan

The Concept-of-Operations, System Requirements, and System Design will be used to develop the software systems. Complementing these documents will be the Operational Design and Implementation Plans. These plans will provide a similar level of detail that describes every aspect of the operational plan beginning with implementation and continuing through maintenance practices and responsibilities. These plans are essentially the operational "playbook" that the partners will use to monitor and manage the system as well as communicate issues and solutions between each other.

Task 6. Construction and Installation

This task recognizes that implementation of the proposed system will require field activities including establishing the new transit service, installation of the DSRC equipment, and installation of the TSP equipment. These activities will be conducted by TDOT and the partners in the grant.

Task 7. System Operational Testing

We are proposing to conduct two discrete types of testing to ensure that the system is robust, secure, and can meet the objectives of the project. First, we will examine each component of the system individually cross-referencing the inputs and outputs with the system requirements and implementation plans. Next, we will conduct a closed pilot test of the entire system for at least a two month period to identify and resolve overall system issues.

Task 8. Operations and Maintenance

There are four subtasks within this overall task corresponding to the initiation of a two-month pilot as a final system operational test, performing system modifications to resolve any issues identified during the pilot, and launch of the operations. TDOT will be responsible for the overall management and oversight of this task, but will heavily leverage the partners as all of the partners will be impacted by the operations.

We expect that the proposed system will become an ongoing component of TDOT's and the partner's daily operations and will therefore operate in perpetuity. However, for the purposes of this grant and for assessing impacts, benefits, and costs, we have assumed a two year operational period for full operations.

Task 9. Data Collection and Evaluation Support

TDOT is committed to evaluation and understanding the relative costs, benefits, and system performance. We will collect data that can be used by USDOT's Independent Evaluator and others to assess performance and benefits/costs. All data collected by TDOT will be scrubbed and then submitted to USDOT's Research Data Exchange at the conclusion of the grant operational period. Annual data deliveries will also be performed to provide data for analysis prior to the conclusion of the project.

Task 10. Communications and Outreach

We expect that this project will generate tremendous interest from States that considering potential approaches for enhancing Traffic Incident prevention and management. We are committed to developing lessons learned and making this information available. At a minimum, the partners will prepare and present findings from the project annually at the Transportation Research Board Annual meeting, the ITS America Annual Meeting, and other similar meetings.

The goal of this project relies upon informing the traveling public of alternative transportation options (transit) as well as unsafe travel conditions (e.g., high probability of a congestion inducing incident). As such, extensive public engagement and outreach will be necessary and conducted. We are tremendously excited about maintaining Tennessee's position as a national leader in technology adoption and will include and feature this project in our public hearings and outreach as a major component for improving the safety and mobility of travelers in the region.

Regulatory Challenges

There are no regulatory challenges that would prevent the installation and adoption of transit signal prioritization along Murfreesboro Pike as this is an extension of an already programmed TIGER grant. Similarly, there are no regulatory issues or challenges associated with the development of Centralized Traffic Management Software and the Predictive Analytics and DSS based upon traffic incidents.

Performance Improvements

Through microsimulation, Noblis found that the "Average network-wide delay was reduced by up to 14% when INC-ZONE was used in addition to an increase in average speeds by up to 8%. There were fewer instances of hard braking in the incident zone with up to 89% reduction in maximum deceleration.1 "We expect similar improvements along I-24. Moreover, because we are creating a DSS based upon predictions of the probability of incidents and taking action based upon those predictions, we expect to see a reduction in the number of incidents due to the staging of State Highway Patrol Troopers, and in clearance times due to staging of TDOT's Incident Response Vehicles. The use of radar and DSRC as a first responder warning system should also reduce the number of injuries and fatalities among first responders at incident zones. Table 4 summarizes the expected performance improvements and identifies the elements of the project that will enable achieving these improvements.

Table 4. Expected Improvements and Performance Measures
Expected Improvement Performance Measure Project Component(s) Contributing to Improvements
1 2 3 4 5 6 7
Safety of Traveling Public
  • Number of traffic incidents on I-24
  • Number of secondary crashes
No Value No Value No Value No Value
Improved Congestion
  • Improved Travel Time Reliability
  • Utilization of Transit Alternatives
  • Transit Travel Time Reliability
  • Number of non reoccurring congestion instances
  • Clearance time of incidents
No Value
First Responder Safety
  • Reduction in number and severity of first responder injuries from vehicle strikes in incident zones
No Value No Value No Value
  1. Centralized Traffic Management Center Software System
  2. Predictive Analytics and Decision Support System for Traffic Incidents
  3. Advanced Traveler Information Systems (SmartWay)
  4. Enhanced Freeway Transit Service with Increased Capacity
  5. Enhanced Transit on Major Arterial Through Expansion of Transit Signal Prioritization
  6. Pre-Incident First Responder Staging Based Upon Predicted Incident Probabilities
  7. DSRC and Radar Based Incident Zone Safety Systems


There are a large number of partners actively participating in this grant as previously discussed. This clearly demonstrates a unified desire by the region to implement technology to improve the safety and mobility of the traveling public as well as to improve the safety of first responders.

Leveraging Existing Infrastructure and Investments

TDOT will leverage previous infrastructure investments to the greatest extent possible in deploying the proposed project elements. Tennessee has long been a leader in proactive forms of congestion management. As an example, high occupancy vehicle (HOV) lanes have been used for over 20 years as a method to decrease congestion on interstates in urban areas in Tennessee, including on I-24 in the Nashville area.

TDOT's SmartWay System provides advanced traveler information on the most heavily traveled highways in the state of Tennessee. Components of this system include four Transportation Management Centers as well as the cameras, radar and video detection systems, roadway traffic sensors, and dynamic message signs that support the flow of information between TDOT and travelers in Tennessee. This project will leverage the SmartWay investment by enhancing the system to include congestion predictions, availability of transit alternatives, and the status of TSM&O activities.

Tennessee Highway Patrol's Crash Reduction Analyzing Statistical History (C.R.A.S.H.) system utilizes probability analysis to allocate their forces at locations where crashes are determined likely to take place in order to reduce the possibility of occurrence. Troopers are currently using this system for staging of patrol units on Interstates and highways throughout the state to reduce speeds and target aggressive drivers. This project will add additional granularity on I-24 segments and enhance first responder staging efforts and safety via DSRC and radar-based incident zone safety systems.

The current Middle Tennessee RTA transit system includes two routes, the 84X and 86X, that provide a freeway express service to passengers traveling between Nashville and Murfreesboro. The 96X provides BRT Lite service between Nashville and Murfreesboro, including routing on Murfreesboro Pike. This project will expand transit signal priority along the 96X route that travels between to deliver improved travel time reliability for transit users. The elements proposed in this project enhance the ongoing efforts of the Nashville MTA/RTA and present an opportunity to test advanced congestion mitigation strategies in a high volume urban corridor setting.

Leveraging Existing ITS Initiatives

TDOT intends to build upon the successes of numerous U.S. DOT ITS initiatives, including the Connected Vehicle, Integrated Corridor Management (ICM), and Active Transportation and Demand Management (ATDM) programs.

This project will leverage the Response, Emergency Staging, Communications, Uniform Management, and Evacuation (R.E.S.C.U.M.E.) bundle of applications. U.S. DOT has previously developed a Concept of Operations, Functional and Performance Requirements, and Prototype System Design documents for this bundle, which was demonstrated in Sykesville, MD. TDOT will seek to build upon these documents in developing its own Concept of Operations and Design documents for the applicable elements of the proposed Next Generation of Integrated Corridor Management through Use of Traffic Incident Predictive Analytics and Decision Support System.

Finally, TDOT will seek to leverage the body of research and strategies utilized in U.S. DOT's ICM and ATDM programs. TDOT has in the past pursued strategies in line with the U.S. DOT ICM program to enhance the development and management of the I-24 Corridor and to build upon the existing deployed infrastructure in conjunction with partner agencies. TDOT will seek to build upon lessons from the ICM demonstrations on US 75 in Dallas, TX and I-15 in San Diego, CA. TDOT will also utilize ongoing ATDM research such as Tools for Tactical Decision-Making/Advancing Methods for Predicting Performance.


TDOT is proposing a four year project duration with an anticipated start date of 1/1/2018 and a grant conclusion date of 12/30/2021. All funds will be allocated to the project prior to the third year of the grant. A proposed schedule in presented in Table 5.

Table 5. Proposed Project Schedule
Task Name Duration Start Finish
1. Project Management 1459 1/1/2018 12/30/2021
Kickoff Meeting 1 1/15/2018 1/15/2018
Monthly Status Reports Monthly 1/15/2018 12/30/2021
Report to Secretary Annually 1/1/2019 12/30/2021
2. Service and Equipment Acquisition 544 1/1/2018 6/29/2019
2.1 Partner Contracts 60 1/1/2018 3/2/2018
2.2 DSRC and Radar Equipment 90 11/21/2018 2/19/2019
2.3 TSP Equipment 135 11/21/2018 4/5/2019
2.4 Transit Vehicles 220 11/21/2018 6/29/2019
3. ConOps and Implementation Planning 324 1/1/2018 11/21/2018
3.1 Concept of Operations 120 1/1/2018 5/1/2018
3.2 System Requirements Specification 111 4/16/2018 8/5/2018
3.3 Architecture and System Design 123 7/21/2018 11/21/2018
4. System Development 589 1/1/2018 8/13/2019
4.1 Central Software 197 1/1/2018 7/17/2018
4.2 PA and DSS 235 11/21/2018 7/14/2019
4.3 SmartWay Modifications 30 7/14/2019 8/13/2019
4.4 Inc-Zone Application 260 11/21/2018 8/8/2019
5. Design and Implementation Plan 159 11/22/2018 4/30/2019
5.1 Implementation Plans 89 11/22/2018 2/19/2019
5.2 Detailed Implementation Designs 70 2/19/2019 4/30/2019
6. Construction and Installation 90 6/29/2019 9/27/2019
6.1 Equipment Installation 57 6/29/2019 8/25/2019
6.2 Software Installation 45 8/13/2019 9/27/2019
7. System Operational Testing 75 9/27/2019 12/11/2019
7.1 Component Testing 30 9/27/2019 10/27/2019
7.2 System Testing 45 10/27/2019 12/11/2019
8. Operations and Maintenance 750 12/11/2019 12/30/2021
8.1 Pilot Launch 60 12/11/2019 2/9/2020
8.2 System Modifications 30 2/9/2020 3/10/2020
8.3 Operational Service 660 3/10/2020 12/30/2021
9. Data Collection and Evaluation Support 1459 1/1/2018 12/30/2021
9.1 System Data 750 12/11/2019 12/30/2021
9.2 Ridership Data 750 12/11/2019 12/30/2021
9.3. Performance Measures 1459 1/1/2018 12/30/2021
10. Communication and Outreach 1459 1/1/2018 12/30/2021
10.1 Stakeholder Engagement 1459 1/1/2018 12/30/2021
10.2 Outreach 544 1/1/2018 6/29/2019
10.3 Knowledge Transfer 360 1/4/2021 12/30/2021

1 Chang, J., Hatcher, G., Hicks D., Schneeberger, J., Staples, B., Sundarajan, S., Vasudevan, M., Wang, P., and Wunderlich, K, "Estimated Benefits of Connected Vehicle Applications: Dynamic Mobility Applications, AERIS, V2I Safety, and Road Weather Management," FHWA-JPO-15-255, August 2015. [ Return to Note 1 ]

Office of Operations