Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
1 Study Objectives This report provides information and guidelines for state departÂ ments of transportation (DOTs) and regional metropolitan planning organizations (MPOs) on updates to modeling and forecasting tools that will be necessary to more appropriately account for the expected impacts of automated vehicles (AVs) and connected vehicles (CVs) on transportation supply, road capacity, and travel demand components. Updates are needed because connected and automated vehicles (CAVs) are expected to prompt disruptive changes to transportation. CAV implementation is likely to influence level of demand, travel modes, planning and investment decisions, physical transportation infrastructure, and geographic areas for all personal mobility and goods movement. Planners and modelers are faced with evaluating public and private investment in roadways and other transportation facilities with only one certainty: disruptive change is on the horizon. Automated technologies in vehicles, efficient communications between vehicles and infrastructure, and a market shift toward economical and flexible shared mobility fleets will transform the current landscape of personal mobility and goods movement. The difficulty of predicting the exact timing, magnitude, type, and locations of the disruptive changes poses new risk for infraÂ structure investment decisions. Both direct and indirect impacts are expected from each of the elements of CAVs, and not all impacts that come from these disruptive technologies will be positive. Experience has shown that indirect and unintended impacts often result from rapid changes, and the planning comÂ munity needs methods to address both potentially positive and potentially negative outcomes. For example, vehicle automation affects the driving task by potentially altering the perceived time inefficiencies related to driving. The idea is that time spent operating a vehicle is wasted, and that time could be spent doing something more productive. Peopleâs value of time will be changed when they are converted from drivers to passengers who will be able to conduct busiÂ ness while in transit to and from a workplace. Another possibility is that the time gained may be spent in leisure activities (e.g., reading, watching TV). The dynamics of this change may result in a shift in the choices people make regarding destination, route, or mode. Likewise, the changes in dynamic travel time as drivers become passengers have implications for tollÂroad modeling and parking costs. Planners and modelers are concerned with longÂrange forecasting of these types of fundaÂ mental, indirect impacts from transportation automation. Change may occur quickly for certain C H A P T E R 1 Introduction Chapter Highlights â¢ Conveys study objectiveâto provide DOTs and MPOs with guidance on updating modeling and forecasting tools related to connected and auto- mated vehicles. â¢ Defines the main challenge addressed by this researchâchanges to the exist- ing forecast modeling paradigm made necessary by the future presence of connected and automated vehicles in traffic streams. â¢ Summarizes report contents and guides readers to specific text on the basis of their information needs.
2 Updating Regional Transportation Planning and Modeling Tools to Address Impacts of Connected and Automated Vehicles modes, while for other modes it may take decades to realize the impacts and obtain market stability. For instance, transit systems may soon be affected as shared rides and comprehenÂ sive mobilityÂasÂaÂservice (MaaS) platforms grow; however, impacts on parking and land use changes may take many years. A significant determining factor in every aspect of CAV adoption and societal effect will be regulatory decision making. The impacts of CAVs on transportation systems will have to be studied and measured as the technology is developed and deployed. However, given what is known today about the potential impact of AV technology when combined with communications systems and sharing behavior, there is clearly a new role for exploratory modeling in a planning context that deals with uncertainty. LongÂrange forecasts are made for 20 to 30 years into the future, so planners can expect that this system of technologies will, by that future time, have a significant impact on the transportation system and travel choices. Some preliminary attempts at modeling have been made with existing tripÂbased and activityÂbased (AB) models, but the results have been somewhat unsatisfying, posing questions instead of answering them. This report describes the need for and the gaps in planning and modeling approaches and tools and presents highÂlevel guidelines for nearÂterm implementations. However, the authors acknowledge that considerÂ ations in the CAV space are changing rapidly and that this report may need updating in the next 3 to 5 years. Defining the Problem: Forecasting Travel Behavior and Technological Changes Forecasting travel and its consequences in urban regions is a difficult prospect, given the scale of the systems and the multitude of influences on travel behavior. How, when, and if traffic conÂ gestion can be reduced is a complicated question because of the uncertainty about many factors. Forecast population growth 30 years into the future can vary significantly, depending on migraÂ tion rates (natural growth is typically stable). Other factors include changes in travel behavior and choices over time, economics, and transportation costs. Potentially radical changes from new technology such as driverless vehicles, a sharing economy, and expanded communications capabilities can all play a part in forecasts as well. In the strict sense, modeling is a mathematical representation of data using formulaic expresÂ sions. Models designed to both predict and test future scenarios can only be as accurate as the forÂ mulas calibrated to match observed data and the forecast independent variables, such as the future number of households. In addition to data, however, the model designâthe structure that defines the independent and dependent variables and the process by which the mobility environment is simulatedâis also critical to the effectiveness of forecasts in providing valuable information to decision makers. This report makes the distinction between model design and modeling (and planning) frameÂ work. A model can be designed to represent information that is reflected in observed data, while a framework is a higherÂlevel, lessÂdetailed presentation of procedures that reflect the anticipaÂ tion of significant changes to data that are collected in the future. A framework is preparatory, while a model design is descriptive. In the past several decades, travel forecasting models have been predicated on the assumption that past trends in travel behavior and choices will continue two or three decades into the future with only minor alterations. This paradigm of transportation modeling was effective because of relative stability being observed in travel behavior over time. Most trips are made by private auto because that mode is widely available, convenient, comfortable, and accessible. People are willing to pay for these travel experience characteristics. In cities where public transportation is readÂ ily available and serves mobility needs more comfortably and economically than private autos, CAV technology is changing the tradi- tional travel forecast- ing paradigm. Not only are changes in planning approaches necessary, but the very structure and com- plexity of models will need to be adjusted as the technologies are deployed. This study was executed under NCHRP 20-102, a task-order support contract addressing critical issues in AV and CV development that DOTs and MPOs are facing. For informa- tion on related studies and reports, see http:// apps.trb.org/cmsfeed/ TRBNetProjectDisplay. asp?ProjectID=3824.
Introduction 3 ridership is robust. These two modes of personal transportationâprivate auto (or walking/ biking) and public transportation [bus or rail, and, to some limited extent, taxis and transportation network companies (TNCs)]âremain the primary mobility choices. Models can be calibrated to observed choices reflected in surveyed data and growth can be applied. This process results in a rational forecast level of demand calibrated to observed data and validated by existing usage levels. The forecast level of demand is then compared with system supply represented by facility capacÂ ity to determine future system performance. The process of calibrating models to observed data, validating the modeling outcomes to existing (or rather, immediate past) conditions, and applying future growth to produce one potential future outcome has been a longÂstanding paradigm in the United States. Recent mobility and technology innovations are prompting a change to the existing forecast modeling paradigm. Telecommunications, vehicleÂ and ridesharing, and robotics have begun, in varying degrees, to change the mobility landscape. With the advent of smartphones, commuÂ nications technology has placed an enormous amount of information about modal availability, routing, system conditions, and other critical transportation information in travelersâ hands. Information communication technologies, through telecommuting and teleshopping over the Internet, are slowly increasing their impact on the need for personal mobility. Personal trip making for access to goods is being supplanted by efficient and economical product delivery. Workplaces are becoming more flexible, allowing many employees to work from home either permanently or as needed. The ease with which someone can gain access to realÂtime informaÂ tion about vehicle location has given rise to TNCs that allow travelers to share a ride and to other services that provide shared use of a vehicle, bicycle, or scooter. These modes are now widely available in many urban areas, but planning and modeling in most regions in the United States have not kept up with the changes. AV technologies are a new element of change expected to influence travel choices. Artificial intelligence, robotics, and connectivity/communications are being incorporated into CAVs. Connectivity is expected to become the norm, with vehicleÂtoÂvehicle (V2V) and vehicleÂtoÂ infrastructure (V2I) components merging with highly automated vehicles (e.g., Levels 3 to 5), creating an informationÂfilled mobility environment. While no wideÂscale deployments of CAVs exist today, this technology is expected to be widely adopted. The technical questions are ones of design, integration with the existing transportation system, and acceptable safety and security. Public policy and liability questions are expected to be more fully addressed when deployment advances. Another aspect of the expected impact on travel of AVs and CVs is auto ownership, availÂ ability, and vehicle use patterns. The ability of AVs to run autonomously (without a driver or passengers) is expected to enhance the utility of TNCs by lowering operating costs for owners. A CAV that runs autonomously and is shared is called a shared autonomous vehicle (SAV). This technology could greatly increase the number of shared vehicles and rides and could change auto ownership patterns significantly. SAVs could be used sequentially or simultaneously (i.e., the pooled versions with high vehicle occupancies). A limitation of the work that has been done to date in modeling CAV use is that it applies imposed or assumed changes in behavior to modeling frameworks. No behavioral data indicatÂ ing trip/tour frequency, length, mode, route, time of day, or other characteristics of AV or CV operations exist, simply because the technology is new and has yet to be deployed. Any changes are assumed by the analysts and may or may not remain applicable when eventually compared with actual deployment and behavioral changes resulting from AV and CV technologies. This report refers to these noncalibrated tests as âmodeling experiments.â Modeling experimentation done by imposition of changes to calibrated parameters and input data (such as trip rates, lengths, and mode choice) to examine the potential impacts of AVs and CVs is testing the sensitivity of
4 Updating Regional Transportation Planning and Modeling Tools to Address Impacts of Connected and Automated Vehicles the models to imposed changes. Existing models do not have AV or CV modes, nor do they reflect behavioral impacts of deployment of AVs and CVs. Modeling parameters, such as the inÂvehicle travel time coefficient in mode choice models, are calibrated to observed conditions. The observed conditions to which these types of parameters are calibrated do not include AV or CV choices for travelers, so these models do not test the actual impacts of AVs and CVs. The needs of MPOs and DOTs may differ when they address the uncertainties posed by future CAV deployment and use. For instance, an MPO or DOT in a highÂgrowth area with significant congestion and investment needs may adopt an aggressive approach to scenario planning and specifically wish to include AVs and CVs and other highÂimpact technological developments in the longÂrange plan for the region. Another MPO or DOT may not have growth issues, so the focus of the longÂrange plan may instead be on economic development and qualityÂofÂlife improvement for citizens. In such a region, the impacts of AVs and CVs could be addressed in an incremental fashion by using only dataÂsupported modeling rather than scenarioÂbased methods. The planning approach that might be chosen has an impact on the type of modeling that is appropriate. An MPO or DOT may decide, as a matter of planning policy, to adopt a range of strategies that best fit the longÂrange requirements a region is addressing. Some regions and states have greater planning capacity to accommodate the additional resources needed to implement changes to planning processes and modeling. While the approach to development of methods may differ between regions and states, all agencies have in common the need for information and guidance on how to plan for and model CAVs. Navigating the Report This report is a resource for understanding and implementing updates in modeling and foreÂ casting tools to more appropriately account for the expected impacts of CAVs on transportation supply, road capacity, and travel demand components. The content is organized in a hierarchical manner. The early chapters present foundational information, and the later chapters present more advanced knowledge that builds on the underlying concepts presented in the early chapÂ ters. The earlier chapters, that is, provide the rationale for accounting for CAVs in planning and modeling activities, as well as pertinent information about technology and regulatory contexts, and summarize uncertainties in benefits and risks. The sub sequent chapters provide highÂlevel guidance on how to practice forwardÂlooking planning and modeling. While CAVs represent new technologies with many moving parts, developing a strategy to begin planning for and modeling CAVs does not have to be complicated. This report provides information about how to get started. The information is geared toward planners and modelers in MPOs and state DOTs of all sizes and geographies. The content is based on reviews of the literature, the professional experience and expertise of the research team, and information gathered in a stakeholder workshop. Table 1 provides an overview of the main sections and corresponding information contained in this report. This report provides information about how state DOTs and MPOs can begin accounting for CAVs in planning and modeling activi- ties. It is intended for agencies both with and without significant resources to undertake new activities.
Introduction 5 Section Description Chapter 1: Introduction â¢ Presents the report purpose, rationale, and organization. Chapter 2: Definitions of CAVs and Current Status â¢ Offers a simple description of CAVs and their enabling technologies. â¢ Defines AVs and CVs. â¢ Describes six levels of AVs. â¢ Summarizes current states of development and deployment for AVs and CVs. Chapter 3: Uncertainties Associated with CAVs â¢ Summarizes uncertainties in CAV adoption timelines and potential benefits and risks. â¢ Describes the uncertainties associated with adoption of the technologies. â¢ Presents a framework of three phases of adoption: (1) testing and early deployments, (2) consumer initial adoption, and (3) system-level organization as CAVs become predominant. â¢ Discusses potential impacts related to safety, congestion, and land development. â¢ Examines critical considerations for planning and modeling in five areas of impact: (1) transportation costs, (2) transportation safety, (3) vehicle operations, (4) electrification (fuel), and (5) personal mobility. Chapter 4: Framework for Planning and Modeling CAVs â¢ Provides a high-level framework for accounting for CAVs in the planning and modeling processes. â¢ Provides a conceptual framework for planning and modeling CAVs. â¢ Discusses the individual elements of the framework: data, planning context, modeling, CAV adoption timeline, and communicating uncertainty. Chapter 5: Planning in the Context of Uncertainty â¢ Discusses approaches for accounting for uncertainty in the planning process. â¢ Reviews how uncertainty is managed in current transportation planning. â¢ Describes the unique challenges in managing uncertainty posed by CAVs. â¢ Identifies methods suited to managing decision making under deep uncertainty: (1) scenario planning, (2) assumption-based planning, (3) robust decision making, (4) info-gap, and (5) dynamic adaptive pathways planning. Chapter 6: Adapting Trip-Based Models to Address CAVs â¢ Provides high-level guidance on accounting for CAVs in trip-based models. â¢ Identifies potential modeling changes. â¢ Discusses the contexts and approaches for (1) land use modeling, (2) auto availability and mobility choices, (3) trip generation, (4) trip distribution, (5) mode choice, and (6) routing and traffic assignment. Chapter 7: Adapting Disaggregate/Dynamic Models to Address CAVs â¢ Provides high-level guidance on accounting for CAVs in AB travel demand models and dynamic traffic assignment methods. â¢ Identifies potential model improvements. â¢ Discusses the modeling contexts and approaches: (1) sociodemographics, (2) land use/built environment, (3) auto ownership/mobility models, (4) activity generation and scheduling, (5) destination/location choice, (6) mode choice, (7) routing and traffic assignment, (8) pricing, and (9) truck and commercial vehicles. Chapter 8: Adapting Strategic Models to Address CAVs â¢ Provides high-level guidance on accounting for CAVs in strategic models developed to supplement more sophisticated modeling efforts as screening tools for evaluating policies. â¢ Depicts the typical strategic model components. â¢ Identifies potential modeling changes. â¢ Discusses the modeling contexts and approaches: (1) sociodemographics, (2) built environment, (3) mobility, (4) accessibility, (5) pricing, (6) travel demand, (7) mode choice, and (8) truck and commercial vehicles. Chapter 9: Communicating in an Uncertain Environment â¢ Provides guidance on how transportation planners and modelers can communicate about the uncertain future. â¢ Distinguishes certainties from uncertainties in a CAV future and presents tips for talking about both. Appendix: Regulatory Context for CAVs â¢ Discusses federal regulatory context for CAVs and state legislation. Table 1. Information in this report.