The Impacts of Vehicle Automation on the Public Transportation Workforce (2022)

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.

5   Introduction Vehicle automation is not a new concept but has recently seen significant technological progress, and there has been significant investment and attention from both the private and public sectors. In the public transit industry, transit vehicle automation has generated substantial research and innovation; however, many challenges remain before full-scale adoption of automated transit vehicles can take place. Although these challenges may be difficult, the interest in automated transit service endures, and progress continues to be made. However, progress toward automated transit service also leads to important ques- tions about the potential public transit workforce effects stemming from transit vehicle automation. To date, the research on potential workforce effects remains quite sparse— due partly to the fledgling state of transit vehicle automation and partly to the significant amount of uncertainty about how and when automated transit services will actually be operated. This research report is the final product for TCRP Project J-11/Task 34, “The Effects of Vehicle Automation on the Public Transportation Workforce,” which had three main objectives: • Identify likely transit automation use cases. • Analyze each use case’s potential effects on the public transportation workforce. • Identify clusters of strategies to prepare the workforce for and mitigate negative effects of transit vehicle automation. In general, this research focused on automated bus transit services, including all fixed- route, flexible-route, and demand-responsive bus, van, and car public transit services. Although rail service can be (and in some cases already is) automated, this research did not investigate rail transit automation. The research dealt only with Level 4 (L4) and Level 5 (L5) auto- mation (see Section 2.1 for a description of the levels of automation). Also, services that Title 49 of the United States Code (U.S.C.) does not define in Section 5302 as public transit were outside of the scope of this research. In 49 U.S.C. §5302, public transportation (or transit) is defined as: Regular, continuing shared-ride surface transportation services that are open to the general public or to a segment of the general public defined by age, disability, or low income; and . . . does not include— • Intercity passenger rail transportation; • Intercity bus service; • Charter bus service; • School bus service; • Sightseeing service; • Courtesy shuttle service for patrons of one or more specific establishments; or • Intra-terminal or intra-facility shuttle services. C H A P T E R 1

6 The Impacts of Vehicle Automation on the Public Transportation Workforce 1.1 Disclaimers This report does not attempt to be an exhaustive and final analysis on either transit vehicle automation or its effects on the transit workforce. This report often uses the words potential, likely, and possible and their synonyms to help readers remember the uncertainty with which statements about transit service automation are made. Readers are cautioned from using the results presented in this report as definitive and are reminded that all discussions and analyses of the future of vehicle automation are replete with assumptions. Understanding the assump- tions made in this study is just as important as correctly interpreting the study’s results. There is much research and analysis underway, and much more needs to be done to fully understand the breadth of possible effects of vehicle automation on the transit workforce (e.g., see the U.S. Department of Transportation’s 2021 publication Driving Automation Systems in Long-Haul Trucking and Bus Transit: Preliminary Analysis of Potential Workforce Impacts). This report is one in a long line of necessary research. However, this report does attempt to help the transit industry understand the range of poten- tial effects on the transit workforce for a limited number of transit automation use cases. The results presented in this report provide the transit industry with discussion points for planning for those potential automation effects on the workforce. 1.2 Key Report Terms There is significant ongoing development in vehicle automation in general, and the automa- tion of transit vehicles is a relatively new area of practice with a growing nomenclature and dictionary of terms. Table 3 provides definitions of key terms as they are used in this report. Of particular importance is the phrase workforce effects. This report uses this phrase repeat- edly as a succinct way to refer to the effects of automated transit services on the transit work- force. Workforce effects could include both job gains or losses and job description changes. 1.3 Organization of this Report This report is broken into the chapters listed and described in Table 4. The report contains six appendices. Table 5 describes each appendix to aid readers in locating desired information. In addition to appendices, this report is accompanied by several attachments, which are downloadable PDF and Excel files, as listed in Table 6. 1.4 Study Methodology No study on vehicle automation can yet claim to have high levels of certainty, and this study is no exception. However, the research team employed a study approach that sought to acknowledge uncertainty in transit vehicle automation while still producing some results for the transit industry that can be useful to understand potential workforce effects of transit vehicle automation. 1.4.1 Conceptual Approach Figure 2 displays the basic conceptual approach of this study. (A task-by-task breakdown of the work performed is provided in Section 1.4.2.)

Introduction 7   Term Definition Affected employees Transit employees who are in a transit job affected by the adoption of automated transit services. Affected transit job A transit job (e.g., bus mechanic) that is affected by the adoption of automated transit services. Automated transit service A transit service that is operated using automated transit vehicles. Automated transit vehicles Transit vehicles (i.e., transit buses, vans, or cars) that are automated at the Society of Automotive Engineers (SAE) L4 or L5. Conventionally driven Vehicles or transit services that are driven by a human operator without automation. May also include vehicles equipped with advanced driver assistance systems not exceeding SAE L3 (e.g., lane-keeping assistance). Directly affected operations jobs The transit jobs that would be most directly affected by the adoption of automated transit services. These jobs include those front-line and supervisory transit jobs that directly operate, supervise, or maintain revenue vehicles. Front-line employee A transit employee whose job is to work on the front lines of transit service. These employees typically include operators and mechanics but could include other non-management positions that work directly with the public or with transit vehicles. Full-size transit bus A transit vehicle, typically used in fixed-route service, usually 30 or more feet long. Full-time equivalents (FTEs) A value that represents the number of full-time positions in a particular job. Full-time positions are counted as one. Part-time positions are counted as 0.5. Indirectly affected key jobs Transit jobs that would be indirectly affected by the adoption of automated transit services. These jobs include those front-line and supervisory transit jobs that work directly with operators, transit planning and scheduling, and vehicle and yard operations and maintenance. These jobs are indirectly affected by automation due to their proximity to and contact with directly affected operations jobs. Job count The number of available positions, counted as FTEs, in a transit job. Job description change A type of effect on the transit workforce; changes to the nature of the work performed for a specific job (i.e., changes in the tasks performed and/or the KSAs required to perform the job) that result from the adoption of a transit automation use case. Job gain or loss A type of effect on the transit workforce; the gain or loss in a transit job resulting from the adoption of automated transit service. Passengers People who are riding or waiting for transit service. Transit A shortened term to refer to public transportation as defined by 49 U.S.C. §5302. May also be referred to as public transit. Transit agency An entity (usually a local or regional governmental or not-for-profit organization) that provides transit service within a defined geographic region. Transit vehicle automation The automation of one or more types of transit vehicles for the purpose of automating a particular type of transit service within a particular operational design domain. Transit vehicle automation refers specifically to SAE L4 or L5 automation in transit vehicles. Transit vehicle automation may not necessarily include automation of non- driving tasks (e.g., fare collection and wheelchair passenger securement). Transit workforce Employees working in the public transit industry at transit agencies and transit-providing private firms. Use case A specific application of a type of automated transit vehicle to a type of transit service in order to provide an automated transit service. For example, automated bus rapid transit is a use case in this report. Also referred to as a transit automation use case. Vehicle automation The automation of vehicles (personal, transit, trucking, etc.) using automation technology at SAE L4 or L5. Workforce effects The effects on the transit workforce resulting from the adoption of automated transit services. Includes job gains or losses and job description changes. Table 3. Key report terms (organized alphabetically).

8 The Impacts of Vehicle Automation on the Public Transportation Workforce Appendix Title Brief Description Appendix A: Description of Methodology to Estimate FTEs per Directly Affected Operations Job Describes how the research team estimated the current number of jobs in each directly affected operations job for each mode and transit agency type. Appendix B: Directly Affected Operations Job Profiles Describes the typical job functions and KSAs for each of the directly affected operations jobs. Appendix C: Indirectly Affected Key Job Profiles Describes the typical job functions and KSAs for each of the indirectly affected operations jobs. Appendix D: Impact Tree Details Provides a graphical representation of the impact trees built to support the workforce effect calculator. There is one tree graphic for each use case and adoption scenario. Appendix E: Concept of Operations for In-Person and Remote Operational Models Describes the research team’s basic concept of operations for both remote and in-person operations of automated transit vehicles. Appendix F: Calculator Assumption Descriptions and Rationale Describes each of the foundational assumptions applied to the workforce effect calculator. Table 5. List of appendices. Chapter # Chapter Title Summary of Contents 1 Introduction • Study purpose. • Report organization. • Study methodology. • Key report terms. 2 Transit Vehicle Automation Overview • Vehicle automation concepts. • Automation-supporting technologies (general and transit specific). • Important considerations related to transit automation, including potential benefits, potential workforce effects, and challenges facing transit automation. 3 Transit Vehicle Automation Use Cases Full descriptions of each of the five transit vehicle automation use cases examined in this research. 4 Job Profiles of Targeted Transit Jobs Descriptions of the transit jobs that were specifically included in this report (i.e., those that will be most affected by transit vehicle automation). 5 Industry Engagement Steps taken by the research team to engage with the transit industry during the course of the study, including results from a front-line employee survey. 6 Workforce Effect Calculator Methodology • Structure and assumptions of the workforce effect calculator (i.e., the tool used to estimate workforce effects due to automation). • Workforce effect calculator inputs and outputs and methodologies for establishing input values. 7 Results: Transit Automation Workforce Effects Potential workforce effects of transit vehicle automation on the transit workforce, including changes in job counts and job duties. 8 Workforce Effects: Suggested guiding principles and strategies to prepare for and mitigate potential workforce effects. Preparation and Mitigation 9 Conclusion • Summary of key findings. • Study limitations. • Recommendations for future research. N/A References Works cited during the conduct of this study. N/A Bibliography Works consulted during the conduct of this study but not cited in this report. N/A Appendices Supporting documents, data, or other materials. Table 4. List of report chapters.

Introduction 9   Foundational to the study were the identication and selection of a limited number of transit automation use cases—specic automated transit services for which automated transit vehi- cles could be conceivably used, and for which there is current progress and signicant interest. (Chapter 3 discusses the ve transit automation use cases and how they were selected.) As with all research in vehicle automation, there was a signicant degree of uncertainty related to when and how automated transit services might be deployed and also uncertainty about what planning and policy decisions transit agencies will make when deploying automated transit services. e research team identied a long list of decisions for each use case that ultimately would impact to what degree each use case would be implemented and how any use case adoption would impact the transit workforce. ese planning and policy decisions were taken to the industry and the research panel through interactive workshops and webinars to collect industry opinions on each decision. Based on industry feedback, the research team could make assumptions about Attachment Title Brief Description Attachment 1: Staffing Count Survey A PDF of the survey the research team used to collect staffing count data from selected transit agencies in the process of estimating current FTEs in directly affected operations jobs. Attachment 2: The American Public Transportation Association’s (APTA’s) APTAtech Workshop Slides A PDF of the slides used at the September 2019 interactive workshop held at the APTAtech: Transportation Technology Conference. Attachment 3: APTAtech Workshop Notes Pages Combined A PDF of the notes pages given to breakout groups at the APTAtech interactive workshop. Attachment 4: Front-Line Survey Instrument A PDF of the online survey instrument used to collect front-line transit employee perceptions of potential benefits and concerns of automated transit vehicles. Attachment 5: Front-Line Survey Recruitment Flyer A PDF of the flyer used to recruit front-line transit employees to complete the front-line survey. The flyer was edited when the survey was extended (not shown). Attachment 6: Industry Webinar Presentation A PDF of the slide deck used during the June 2020 industry webinars. Attachment 7: Industry Poll Data An Excel workbook containing a summary of data received through poll questions administered during the June 2020 industry webinars. Attachment 8: Task Impact Ratios An Excel workbook containing current task lists and estimated potential task percentage changes per directly affected operations job and use case. Attachment 9: Workforce Effect Estimates for Directly Affected Jobs An Excel workbook containing the full details of estimated potential workforce effects by use case, operational model, adoption scenario, agency type, and affected job. Table 6. Table of attachments. Identify Transit Automation Use Cases Make Planning and Policy Assumptions Develop Workforce Effect Calculator Identify Preparation Strategies Complete Final Report Figure 2. Conceptual approach to the study.

10 The Impacts of Vehicle Automation on the Public Transportation Workforce planning and policy decisions and then incorporate those assumptions into the workforce effect calculator. The workforce effect calculator provided a tool by which the research team could estimate the number of changed jobs and job gains or losses that might result from the adop- tion of each use case. Based on those estimated effects, the research team identified potential preparation strategies that could be customized based on the use case and on local needs. Last, the research team compiled the result of this work into this final report. 1.4.2 Task-by-Task Methodology The research team accomplished the conceptual approach shown in Figure 2 through the execution of the eight research tasks discussed in the following sections. 1.4.2.1 Task 1: Identifying and Detailing Transit Vehicle Automation Use Cases 1.4.2.1.1 Purpose. The researchers identified the most likely transit automation use cases in the next few decades and described the potential operational impacts, enabling technologies, and any planning and policy decisions that would significantly moderate the effects on the transit workforce. 1.4.2.1.2 Activities. The research team reviewed the academic and professional literature, as well as federally supported research and strategic documents—particularly the Federal Transit Administration’s Strategic Transit Automation Research Plan (STAR Plan) (Machek et al. 2018) to identify the most likely transit automation use cases that should be included in this research. (The field of transit automation is changing rapidly and, as of the writing of this report, some of the timelines in the STAR Plan may be now out of date. The research team did not adjust or change any of the FTA’s original STAR Plan timelines.) The resulting use cases are described in detail in Chapter 3 of this report but are listed as follows: • Bus automation for maintenance and yard operations. • Low-speed automated shuttles. • Automated bus rapid transit. • Automated mobility on demand. • Automated local bus service. 1.4.2.2 Task 2: Learning from Elsewhere: Reviewing Historical and Non-transit Automation Workforce Effects 1.4.2.2.1 Purpose. The researchers used previous transitions from manual to automated work to learn about potential workforce effects of transit automation. 1.4.2.2.2 Activities. The research team reviewed other industries’ automation histories, including agriculture, warehousing, manufacturing, banking, retail, containerization in the ship- ping industry, and aviation. The research team also reviewed existing research on the potential workforce effects of automated vehicles (AVs)—including trucking, transit, and taxis—as well as the automation of personal vehicles. The research team incorporated the lessons learned from this research into its work on the remaining tasks. 1.4.2.3 Task 3: Detailing the Transit Labor Market and Developing the Workforce Effect Calculator Framework 1.4.2.3.1 Purpose. The research team established a taxonomy of transit jobs that were likely to be affected by the identified use cases. The researchers described the workforce currently in these jobs (including employee counts and demographic characteristics, if available) and

Introduction 11   developed typical task lists for each job. The team created a framework for the workforce effect calculator to model the number of automation-effected employees in each job—disaggregated by bus transit mode and agency type (i.e., rural, small urban, and large urban). 1.4.2.3.2 Activities. The research team: • Compiled and synthesized transit job titles in industry publications and identified the transit jobs that would be most affected by transit vehicle automation, both directly and indirectly. • Prepared a demographic analysis of likely affected transit jobs using existing data, if avail- able. Highlighted issues related to workforce shortages, current working conditions, and any disproportionate representation of minority or low-income populations. • Identified a list of tasks that typically fall within each affected transit job. Used existing data sources, research team expertise, and research panel feedback to estimate the percentage of time associated with each task. • Determined the number of employees currently in each affected transit job using both existing data [e.g., the National Transit Database (NTD)] and a survey of transit agencies. • Created a workforce effect calculator framework (or conceptual design) to estimate the number of jobs that would be affected by each use case. • Identified key assumptions in the workforce effect calculator framework that would be affected by planning and policy decisions that need confirmation from the transit industry. For example, would transit agencies implementing automated services that produced cost savings likely reinvest those cost savings in increased service or use the cost savings to reduce the amount of needed subsidy? (These planning and policy decisions were presented to the industry for feedback in Task 5.) The results from Task 3 are discussed in various sections of this report: • Affected transit jobs and their basic characteristics are discussed in Chapter 4. • The number of current employees in impacted transit jobs and the methodology used to estimate the number of current employees are discussed in Chapter 6 and Appendix A. 1.4.2.4 Task 4: Producing the Interim Report 1.4.2.4.1 Purpose. The interim report summarized the transit automation use cases, presented the profiles of affected transit jobs, described the workforce effect calculator frame- work, and identified planning and policy decisions that require industry confirmation. 1.4.2.4.2 Activities. The research team wrote the interim report and submitted it to the research panel for review and comment. Panel comments were documented for inclusion in the final report. 1.4.2.5 Task 5: Engaging the Industry 1.4.2.5.1 Purpose. The research team presented transit industry stakeholders with initial thoughts about transit automation use cases and likely affected transit jobs and obtained feed- back on planning and policy decisions to help build assumptions into the workforce effect cal- culator. The team also collected perceptions of automated transit services from front-line transit workers. (The activities under Task 5 were conducted at various points throughout the study.) 1.4.2.5.2 Activities. Industry engagement fell into three main activities, as follows, in chronological order: • An in-person workshop at the APTAtech: Transportation Technology Conference in September 2019 that had 21 attendees from several different organizations, including

12 The Impacts of Vehicle Automation on the Public Transportation Workforce transit agencies, AV firms, and a representative from a transit union. The in-person workshop sought to collect feedback from the transit industry about key planning and policy decisions that would have implications on workforce effects of transit vehicle automation. • A survey of front-line transit employees conducted between October and December of 2019 in order to understand the perceptions of front-line transit employees regarding transit vehicle automation. The survey received 129 valid responses from bus operators, bus mechanics, and other transit personnel. • Interactive industry webinars held on June 16 and 18, 2020, in order to collect quantitative data on the key planning and policy decisions. The webinars were attended by 89 participants. Chapter 5 contains the details of these activities, including descriptions of the methodology and results. 1.4.2.6 Task 6: Assessing Transit Vehicle Automation Workforce Effects 1.4.2.6.1 Purpose. The researchers finalized the workforce effect calculator and then, for each transit automation use case, estimated and documented the potential workforce effects given the assumptions about planning and policy decisions, supported by industry outreach in Task 5. Workforce effects included both job gains or losses (changes in the number of available positions) and job description changes (changes in the characteristics of an existing job). 1.4.2.6.2 Activities. The research team finalized the design of the workforce effect cal- culator (incorporating industry feedback on the planning and policy assumptions) and then used the calculator to estimate the job gains or losses of each use case for each affected transit job. The team presented preliminary findings to the research panel and then revised assump- tions and scenarios based on panel feedback to calculate the final estimated workforce effects. Potential job description changes were also documented and discussed with the research panel, including, for example, potential changes in job tasks and changes in transit service delivery models. 1.4.2.7 Task 7: Identifying Preparatory Strategies to Maximize Positive Effects and Minimize Negative Impacts on the Workforce 1.4.2.7.1 Purpose. After identifying workforce effects, the researchers identified recom- mended preparatory strategies based on industry best practices. 1.4.2.7.2 Activities. The research team: • Reviewed existing strategies for responding to automation-related workforce effects in the transit industry, government, and other industries. • Investigated best practices from past research on transit workforce development. • Identified what workforce development needs exist to prepare for potential workforce effects. • Made recommendations for strategies to prepare the transit workforce for the transition to automation. Chapter 8 contains the analysis and recommendations from Task 7. 1.4.2.8 Task 8: Producing the Final Deliverables This task was dedicated to producing the final deliverables from this research. All deliverables from this research, including the final report and supporting appendices, were reviewed and commented on by the research panel.