Realistic Timing Estimates for Automated Vehicle Implementation (2023)

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   Review of Existing Forecasts Automated vehicles (AVs) have become a huge topic of interest over the last decade. Initial searches for existing AV forecasts found that there are dozens, if not hundreds, of publications that discuss the operation of AVs in some potential future. Most of the litera- ture that discusses the future effects of AVs assumes that AVs will become the dominant form of transportation and then explores how the effects of AVs may affect transportation systems and society more broadly. However, relatively few publications provide details meant to support assumptions that significant AV deployment is inevitable. The quality of fore- casting and the types of assumptions made vary widely. The research team found 16 publications that provided explanations of AV deployment forecasting methods and were considered worth summarizing. An extensive review analyzed these forecasts and concluded that, with no substantive AV market in place today, the current framework for AV forecasting relies entirely on subjective factors. Key observations also included recognition that there was a wide variety of approaches, variability in time-horizons and geographical scope, and, in most instances, a “cumulative bundling” in the use cases of interest: some focused on robotaxis, others on privately operated AVs, and most on a combination of several scenarios. Eight of the 16 forecasts were published by financial institutions (some of which may have investments in the auto- motive industry). Despite the deficiencies noted, the research team identified and documented the influ- encing factors that others have used in their approaches to ensure that any potential factors were not omitted. This process allowed the research team to better understand the context in which various factors were used and the range of forecasting approaches taken by others. A Forward-Looking Framework The intent of this project was to deliver a framework of considerations for decision- making focused on vehicle automation and a set of assumptions and tools that agencies can use regardless of their size, geographic location, miles of roadway, number of transit vehicles, or funding sources. The framework proposed as a result of the research effort can best be described in terms of three basic components or inputs that are ingested by the model and used to determine output scenarios. This process is outlined in Figure 1. To better understand AV adoption, it is important to look beyond the umbrella term “automated vehicle” and examine specific uses for automation. As AVs are adopted, there are likely to be independent adoption rates and influencing factors for different use cases, S U M M A R Y Realistic Timing Estimates for Automated Vehicle Implementation

2 Realistic Timing Estimates for Automated Vehicle Implementation depending on what becomes safely possible while addressing a viable market need. The proposed use cases are as follows: • Freight Vehicles • Transit Vehicles • Fleet Vehicles (ride-hailing robotaxis) • Personal Passenger Vehicles Classifying AV deployment according to these four use cases constitutes the first of three inputs of a proposed framework for understanding how infrastructure owners and operators (IOOs) should expect AVs to be implemented over time. Input parameters for and output results of the eventual models (premarket and market versions) will vary depend- ing on each use case—as will the operational design domains (ODDs) covered. The second input for the proposed framework reflects a deeper analysis of the informa- tion obtained through the literature review. From the beginning, this project has focused on defining input parameters that may be suitable for estimating deployment timeframes and the degree of certainty of these estimates. IOOs would enter these input parameters, along with other information, into a computer model (premarket or market) and use the resulting outputs to guide planning processes. The research team, working with the project panel, developed the following six categories of input parameters: Use Cases - Freight Vehicles - Transit Vehicles - Fleet Vehicles - Personal Passenger Vehicles Key Actors - Vehicle Producers and Providers - Infrastructure Owners and Operators - Private Transportation Operators - Financial Institutions and Investors - Consulting and Strategy Firms and Researchers - Public Officials and Politicians - Consumers Input Parameters - Current Technological Capabilities - Government Regulations - Average Travel Behavior and Comparable Benefits by AV and Non-AV - General Consumer Trends - Technology Costs - Investor Sentiment and Forecasts Scenarios - Trigger Points - Considerations - Critical Paths - Confidence Intervals Model Figure 1. AV adoption timeline estimation model framework: Inputs and outputs.

Summary 3   • Current Technological Capabilities (e.g., number of Level 4 AVs deployed, number of crewless Level 4 AVs deployed); • Government Regulations (e.g., whether the U.S. Congress has passed AV legislation, what this legislation entails, what AV policies have been adopted at a state level); • Average Travel Behavior and Comparable Benefits by AV and Non-AV (e.g., traveler’s value of time spent as a driver vs. as a passenger, consumer’s willingness to rideshare and vehicle share); • General Consumer Trends (e.g., media coverage of AV successes and failures, consumer acceptance of the capabilities of partially automated vehicles); • Technology Costs (e.g., AV hardware and software capital costs, AV operating costs); and • Investor Sentiment and Forecasts (e.g., updated industry forecasts, levels of technology investment and subsidy). Each of the six categories encompasses a range of potential values and, in conjunction with the other input conditions (i.e., use cases and key actors) will result in forecasts having various trigger points, key milestones, and prioritization. In some instances, possible values currently exist, in others, there will be an evolution as the AV industry transitions from premarket to market posture. The third input of the proposed framework encompasses the key actors and decision- makers in the AV space and considers what levers such individuals have at their disposal. The seven identified key actor groups are typically directly involved in producing AVs, developing AV technology, influencing how AV technology is used, or having a direct role in using AV technology and are as follows: • Vehicle Producers and Providers, • Infrastructure Owners and Operators, • Private Transportation Operators, • Financial Institutions and Investors, • Consulting and Strategy Firms and Researchers, • Public Officials and Politicians, and • Consumers. Future Directions Ideally, going forward practitioners will develop the actual model and then test it using the inputs proposed herein. Beyond initial development, the model should continue to be updated when certain levels of market maturity are reached, and eventually, efforts should transition to the market model. Practitioners can begin to create the market model while developing the premarket model and then transition to the market model once fewer uncertainties remain. As the proposed framework is being developed into a computer model, the transporta- tion community will need to spend resources on developing more metrics and collecting objective data. Such data will enable the transportation community to enhance current models (or build new ones).