Mobility on Demand and Automated Driving Systems: A Framework for Public-Sector Assessment (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.

45 REPORT CHAPTER 6: CONNECTED AND AUTOMATED VEHICLES CONNECTED AND AUTOMATED VEHICLES SECTION OVERVIEW This section discusses connected and automated vehicles and answers key questions including: • What are connected and AVs? • What factors will influence AV deployment? • What are the potential impacts of AV deployment? • What current pilots for AVs exist? WHAT ARE CONNECTED AND AUTOMATED VEHICLES? Innovations in shared mobility, transportation commodification, digital information and fare payment integration, and electrification are supporting the development of connected and automated vehicles (CAVs). CAVs are equipped with communication technologies that allow them to relay information to and from the driver, other vehicles (i.e., vehicle-to-vehicle [V2V]), roadside infrastructure (i.e., vehicle-to-infrastructure [V2I]), and to cloud-based data storage platforms. While all CAVs are AVs, not all AVs are equipped with the technologies that allow them to be classified as a CAV. In turn, vehicle automation could potentially create new opportunities for these advancements. Currently, there is no standard definition for SAVs due to differences in technical opinion about whether the vehicle or the driving system is shared. At present, terminology to define SAVs include SAV, shared automated driving system (ADS)-dedicated, and shared-ADS-equipped. Despite the variety of terms defining SAVs, experts generally agree upon vehicle automation levels. SAE International, a global standards organization for mobility engineering, has developed six levels of automation (Level 0 to 5) to define the level of control that is needed from the human operator or provided by the vehicle. SAE’s levels of automation are: • Level 0: These vehicles are not automated, and drivers perform all the tasks. • Level 1: These vehicles automate only one primary control function. • Level 2: These vehicles have automated systems with full control of specific vehicle functions such as accelerating, braking, and steering. Drivers must still monitor driving and be prepared to immediately resume control at any time. • Level 3: These vehicles allow drivers to engage in non-driving tasks for a limited time. Vehicles will handle situations requiring an immediate response; however, drivers must be prepared to intervene within a limited amount of time when prompted. • Level 4: A human operator does not need to control the vehicle, if the vehicle is operating under the specific conditions that it is intended to function. • Level 5: Vehicles are capable of driving in all environments without human control.

46 REPORT At Level 4 automation, vehicles can operate without an onboard human driver within specific conditions (e.g., road speed, geography, climate). At present, companies are working to develop AVs that can operate at Level 4 automation. Figure 11 summarizes these six levels. Figure 11. SAE International Automation Levels Source: SAE International, 2019 Vehicle automation, using both AVs and CAVs, has the potential to create new opportunities for public transportation including cost savings; automated pick-up, drop-off, and charging; and more economical and convenient demand-responsive services. Increasing automation is enabled by technologies that the AV industry is continuously developing and refining. WHAT FACTORS SAFELY ENABLE AVS? The safe, successful deployment of AVs can be enabled through a set of influential interrelated factors—sensors and perception algorithms, mapping technology, operational design domain (ODD)9, connectivity, physical infrastructure, and human factors. Sensors enable the functionality that ultimately determines the ODD and what is possible for MOD and AVs, such 9 The ODD consists of the specific conditions that AVs are designed to operate in. These conditions include geography (e.g., within a certain zip code or radius); traffic (e.g., whether pedestrians and/or other vehicles are present); speed; and roadway typology (city streets, closed communities, highways).

47 REPORT as where they can operate, how much they cost, and what the passenger experience is like. Before AV transportation services can be deployed, organizations must solve the technical challenges associated with AVs sensing, planning, and acting safely in diverse and unpredictable driving scenarios. They must also provide evidence that the system is safe and will operate as intended. Figure 12 depicts the typical sensors and technologies expected in AVs. Figure 12. AV Enabling Technologies Source: Center for Sustainable Systems, 2017 As sensors, software, mapping, and infrastructure improve, AV technologies can be deployed more widely to support new use cases and business models. The following technologies will enable AVs: • Systems engineering: AV systems and equipment must be integrated in a rigorous and adaptable approach that allows them to be functional, fixable, and safe. • Sensors and perception: Sensors gather information about AV surroundings. Perception algorithms then process and synthesize this information to develop an operational map. • Mapping: AVs require a priori and “almost” real-time mapping technologies for safe operation. A priori maps provide relatively static information about elements, such as street geometry, while maps that offer information almost in real time can provide positioning information dynamically. • Position, navigation, and timing (PNT): PNT can allow vehicles to make safety-critical decisions based on vehicle positioning in the surrounding environment and in accordance with accurate timing and coordination with other internal systems.

48 REPORT • Communications and security: A variety of communication systems and networks can provide robust, reliable, and secure information to and from AVs and surrounding vehicles and infrastructure. • Human factors: Despite high levels of automation, AVs will need to be able to easily transition between human and vehicle control. • Dedicated short-range communications (DSRC): DSRC is a two-way wireless communication capability that permits very high data transmission and can be used by vehicles to communicate with other vehicles, infrastructure, and personal mobile devices (e.g., cell phones) directly. Some auto manufacturers are now considering the use of C- V2X communications instead of DSRC. The more recent cellular vehicle-to-everything (C-V2X) capability has the same purpose as DSRC of direct communication link between vehicles, infrastructure, and personal mobile devices. The main difference is that C-V2X is based on cellular modem technology. Appendix B: Enabling Technologies provides further information on these enabling technologies. AV Features and Vehicle Design Technology enables unique AV functionalities, which vehicle manufacturers or companies market as features. AV features determine which use cases a vehicle can support and may influence the design of other vehicle attributes, such as the powertrain and cabin size. Enabling technologies may also allow AVs to be used for functions besides passenger mobility (e.g., goods delivery). SAE defines an AV feature as “a driving automation system’s design-specific functionality at a [different] level of driving automation within a particular ODD.” Basic AV functions and operations are enabled by internal systems that combine information from a variety of sensors (e.g., odometry, GPS) to identify and operate on different navigation paths. In addition to these standard internal systems, AVs may have a variety of additional features. Figure 13 forecasts the deployment of these features, with the rate of deployment dependent on driving complexity and vehicle velocity.

49 REPORT Figure 13. Forecasted Timelines for AV Feature Deployment Source: Dokic, Muller and Meyer, 2015 The research team reviewed over 50 literature sources including marketing releases, concept vehicles, technical journals, and conference proceedings and then identified seven categories of AV features (SAE Level 3 through 5). Table 10 summarizes the seven categories of AV features. These features will be deployed on different timelines that are driven by factors, such as technical difficulty (e.g., more challenging).

50 REPORT Table 10. AV Feature by Category Category Generic Feature (SAE Levels) Select Examples of Feature 1 L3 Conditional Automated Traffic Jam Drive Audi Traffic Jam Pilot 2 L3 Conditional Automated Highway Drive Mercedes Highway Pilot Truck 3 L4 Highly Automated Low-Speed Shuttle Auro Self Driving Shuttle, CityMobil2 Automated Shuttle, EZ10 Self Driving Shuttle, NAVYA ARMA Shuttle, Olli Local Motors Shuttle, Varden Labs Self Driving Shuttles, Smart vision EQ fortwo 4 L4 Highly Automated Valet Parking Bosch Valet Parking 5 L4 Highly Automated Emergency-Take Over Toyota Guardian 6 L4 Highly Automated Highway Drive Audi Highway Pilot, Bosch Highway Pilot, Otto Trucking Highway Assist, Volvo Highway Assist, Ike Robotics, Starsky Robotics, Waymo 7 L4 Highly Automated Vehicle/TNC (Robo Taxi) Audi Elaine, Audi Aicon Baidu Automated TNC, Bolt EV platform, GM Cruise Automation TNC, Waymo Automated TNC, Honda Automated Drive, Mercedes-Benz F 015, NIO EVE, Nissan Autonomous Drive, Rolls Royce 103EX, Tesla Self-Drive, Uber Automated TNC, Volkswagen I.D. Pilot, Volkswagen Sedric, Volvo Intellisafe Auto Pilot, Ford Automated TNC, Toyota Chauffeur, BotRide Automated TNC Two types of generic AV features are most likely to support MOD and SAV applications. These features include L4 low-speed shuttles and L4 TNCs. Other AV features, which are often marketed as privately owned and operated vehicles, typically do not support MOD applications. L4 Low-Speed Shuttles SAE L4 low-speed shuttles are AV shuttles that drive along a predetermined route. Currently, this feature operates on a fixed route within a geofenced area, such as a retirement community or school campus. This feature is expected to operate in urban areas, such as central business districts, and may frequently interact with other roadway users including pedestrians and bicyclists. The system does not need an onboard-driver control interface (e.g., conventional steering wheel, brake, and accelerator pedal). The shuttle is often limited to speeds below 25 miles per hour (mph). An example is Olli, an EV shuttle manufactured by Local Motors. Olli vehicles are being tested in several locations including Phoenix, Arizona and National Harbor, Maryland, as well as in Germany (Local Motors, 2018). Olli uses IBM’s AI-powered Watson technology to interactively communicate with onboard passengers and can remain fully functional even with the loss of GPS. Olli can be part of a fleet management system due to a central operation center designed to solve the transportation needs of large campuses and municipalities. A smartphone application is available for users to find existing routes, share a ride, and input pick-up and drop-off locations for door-to-door service.

51 REPORT In October 2015, NAVYA launched ARMA, an electric AV shuttle. The NAVYA ARMA does not require a driver or specific infrastructure for operation. It can avoid static and dynamic obstacles, transport up to 15 passengers, and safely drive up to 28 mph. Its batteries are recharged by induction and last from five to 13 hours, depending on the configuration, route specifications, and traffic conditions. The most recent version of the NAVYA ARMA includes 3D-route mapping (NAVYA, n.d.). EasyMile is a startup specializing in providing the software powering AVs and last-mile smart mobility strategies. The company’s EZ10 model is an electric shuttle designed to cover short distances and predefined routes in multi-use environments. EZ10 can operate in different modes, depending on the amount of supporting infrastructure, and can operate up to 25 mph. The shuttle service runs on virtual tracks that an operator can easily configure to accommodate sudden shifts in demand. The vehicle’s hybrid sensing approach combines shuttle localization through vision, laser, and differential GPS data. Fleet management software enables remote and real-time monitoring and control of the fleet of EZ10 shuttles (EasyMile, n.d.). L4 Robo Taxis L4 robo taxis are highly automated driverless vehicles anticipated to be operated primarily by TNCs. The ADS feature allows the vehicle to operate without the need of a driver on a predefined set of roads or geographic area and within the specified ODD. The ADS automatically collects and processes data from onboard sensors to perceive surroundings, such as obstacles and relevant signage, and identify the appropriate navigation trajectories. Robo taxis are capable of navigating through the identified trajectories safely with no human input. This feature is expected to operate in urban areas, such as central business districts, and may frequently interact with other roadway users (e.g., pedestrians and bicyclists). For this feature, the ADS is responsible for the dynamic driving task (DDT) fallbacks and for achieving a minimal risk condition. Minimal risk conditions could include turning on the hazard flashers, maneuvering the vehicle to the road shoulder and parking it, and automatically summoning emergency assistance, which concludes the fallback response. However, robo taxi vehicles that are designed to also accommodate driver operation (whether conventional or remote) may allow a user to perform the DDT fallback if they choose to do so. If a driver is allowed to handle the DDT fallback and accepted, they may simply continue driving manually instead of achieving a minimal risk condition. However, an L4 robo taxi does not need to be designed to allow a user to perform DDT fallback and may disallow this ability (SAE, 2018). During the early deployment stages of L4 robo taxis, the car manufacturers will primarily target TNCs with large fleets. Large fleet companies that are able to afford the highly AVs (HAV) at early deployment and could be the key players for expanding the HAV market by using these vehicles for a variety of purposes. L4 HAV features could be used in shared mobility scenarios, such as by an individual who decides to share their personal vehicle or by shared mobility companies using a fleet of AVs for MOD purposes. Examples of this feature include the Google Car (Waymo, 2017), Tesla Self- Drive (Tesla, 2017), Volkswagen I.D. Pilot Mode, Volvo IntelliSafe Auto Pilot (Volvo, 2017), and Nissan Autonomous Drive. Figure 14 shows an example of an L4 HAV concept vehicle.

52 REPORT Figure 14. Example AV Concept Vehicle Source: Audi Aicon Concept Car—Autonomous on Course for the Future, 2018 The interior of these vehicles is being reimagined without the need to accommodate a driver. This allows manufacturers to eliminate steering wheels and reconsider seat directionality and function. The interior design of AVs may influence the transportation experience including enabling passengers to work and engage in multimedia. Long commutes could become a time for productivity and entertainment. However, these new designs may face regulatory barriers as they may render the vehicle unable to pass current Federal Motor Vehicle Safety Standards which currently require vehicle controls (e.g., steering wheels, gear shifts) to be located within reach of the vehicle operator. The National Highway Traffic Safety Administration (NHTSA) is researching these challenges (NHTSA, 2018). However, NHTSA may not begin developing rules on different seating positions in passenger cars until late 2020, and it could take until 2025 before NHTSA makes comprehensive changes to vehicle safety standards relating to AVs (Anderson, 2019). WHAT FACTORS WILL INFLUENCE AV DEPLOYMENT? As private industry and other organizations continue to develop technologies and features that allow for AV operations, three key areas could maximize the deployment of AVs: • Business models, use cases, and partnerships: Role of business models, use cases, and partnerships in enabling AV pilots and programs; • Built environment: Relationship between AVs and the built environment, land use, and infrastructure; and • Policies and regulations: Key policy and regulatory considerations that could impact AVs include the role of regulation, who regulates, what is being regulated, and who is being regulated. Business Models, Use Cases, and Partnerships The following subsections include a review of potential MOD and AV business models, use cases, and partnerships.

53 REPORT Business Models Stocker and Shaheen (2018) review potential future SAV business models and predict that they could evolve from MOD business models currently available today including B2C, P2P, and for- hire (see the “Business Models” subsection of Section 2: Shared Mobility for further information.) As AVs become more common, the B2C and P2P business models may converge with for-hire services (e.g., carsharing and TNC service models converge as vehicles pick up riders on- demand). B2C may also expand to include automated delivery of goods, such as food. Over time, B2C models that include goods delivery could evolve to P2P business models similar to how some services (e.g., UberEats) operate today. Because a SAV network at scale will not require a human driver, there will no longer be a need to distinguish between “for-hire” and other business models. Instead, the most important distinguishing factor may be determining who owns the AV and who owns the network or platform on which the vehicles are shared. Similar to MOD, there are a variety of organizational responsibilities associated with SAVs including booking, routing, payment, data collection, user access, vehicle maintenance, and providing insurance. One or more entities or individuals could manage these responsibilities. Public agencies may also play a role in developing partnerships to fulfill these responsibilities, performing these responsibilities themselves, and regulating stakeholders in these roles. Section 1 of the Shared Automated Vehicle Toolkit provides further information on a variety of stakeholders’ roles regarding labor in different MOD and AV business models. In the future, SAVs will likely be deployed in one of three business models: (1) B2C, (2) P2P, or (3) hybrid. A hybrid model includes SAVs comprised of both B2C and P2P vehicles. In this model, any combination of individuals, businesses, nonprofits, or public agencies could own, administer, or manage aspects of the network. Table 11 shows a chart of potential SAV business models (Stocker and Shaheen, 2018).

54 REPORT Table 11. SAV Business Model Examples SAV Business Model Title Description Current non-SAV Example B2C Single Owner- Operator Would employ a SAV fleet that is both owned and operated by the same organization B2C carsharing operator (e.g., Zipcar, car2go) both owns and operates a SAV fleet Different Entities Owning and Operating Two (or more) companies partner to provide SAV services Partnerships between rental car and TNC companies are one example of this model (e.g., Avis partnership with Lyft) (CNN, 2018) P2P Third-Party Operator A third-party operator would control network operations of a P2P fleet, likely taking some monetary contribution from the vehicle owner, user, or both in exchange for their services P2P carsharing or TNC services, but where many vehicles on the network are fully automated Decentralized P2P Operations Individually owned AVs where operational aspects are not controlled by any one centralized third party and are instead decided upon by individual owners and agreed upon operating procedures, possibly facilitated by emerging technologies, like blockchain Arcade City, an Austin-based TNC service that operates P2P services with no central intermediary Hybrid Multiple Entities Owning, One Entity Operating An entity that owns a portion of the SAVs in their fleet but also includes individually owned AVs that join the entity’s shared fleet when individuals make their vehicles available for sharing on the network TNC mixed-ownership fleet Different Entities Owning, Same Entity Operating A third party that does not own SAVs themselves but brings online individually owned and entity-owned AVs on a shared network of vehicles that they operate Getaround (P2P and B2C carsharing) Source: Stocker and Shaheen, 2018 Use Cases These business models may inform different SAV use cases. Operators may have to consider different use cases depending on the AV technologies and SAV vehicle characteristics deployed (e.g., vehicle speed, passenger capacity, and propulsion type). The following SAV use cases were developed through a literature review and expert interviews. These use case scenarios were then augmented with feedback from a stakeholder engagement session. Appendix A: Stakeholder Engagement summarizes the session findings and participants, and Section 2: Shared Mobility contains background information on some of these use cases.

55 REPORT Supplementing Public Transit Services Key characteristics of public transit services for travelers are social equity, safety, courtesy, and reliability. During the COVID-19 pandemic and recovery, public health and public trust in public transit systems may also be defining characteristics of these services. SAVs may be able to enhance existing public transit services by increasing perceptions of safety, courtesy, and reliability. The following five case studies detail how SAVs can support public transit operations. In addition, Appendix C: SAV Pilots provides further information on current use cases and pilot projects. 1. Paratransit services: Paratransit has historically been costly to operate, and SAVs may enable cost reduction by removing the human driver. However, paratransit users may require a human attendant and/or assistive technologies (e.g., robotic arms, voice-enabled controls) to assist individuals onto the vehicle, help secure wheelchairs, or provide other services. During a stakeholder feedback session, an attendant noted that individuals with mobility disabilities are concerned about assistance in boarding and alighting vehicles. On the other hand, other individuals with disabilities are excited about the potential of MOD and AV technologies. In addition to cost savings, SAVs may also be able to address paratransit service challenges, such as long wait times. If SAVs are equipped with more efficient routing and communication systems than current paratransit systems use, SAVs may be able to decrease wait and trip times for paratransit users. While the research team was unable to document any operational SAV paratransit services at present, the Jacksonville Transportation Authority (JTA), in partnership with NAVYA, began testing Americans with Disabilities Act (ADA)-accessible AV shuttles on a test track in 2019 (Metro, 2019). At Texas A&M University, researchers are developing protocols and algorithms that allow people with disabilities and SAVs to communicate with words, sound, and electronic displays to help meet accessibility needs (Saripalli, 2017). 2. First-mile and last-mile connectivity to public transportation: SAVs may be able to help bridge first- and last-mile gaps by offering connections to public transit stops and stations. For example, the Optimus Ride pilot transports employees of businesses in the South Boston Waterfront district to public transit (New Urban Mechanics, 2018). In another pilot, the Contra Costa Transportation Authority (CCTA) is using automated shuttles to connect Bishop Ranch shopping center workers in San Ramon, California, to Bay Area Rapid Transit (BART) (Bloom, 2018). 3. Fixed-route public transportation: SAVs could operate a fixed-route service, similar to a bus route or monorail system. Operational deployments of fixed-route SAVs may be simpler for initial deployments because engineers do not have to account for the variables in a more dynamic or on-demand system (Stocker and Shaheen, 2018). Separating AVs to remove the possibility of interacting with human drivers could allow for the vehicles to operate at much greater speeds and improve safety. In June 2019, EasyMile started to transport Texas Southern University students and staff along a fixed route with three stops around the university campus (EasyMile, 2019). 4. Low-density/spatial gap-filling services: SAVs could provide on-demand or fixed-route service to travelers in low-density areas that may lack other transportation options or only have access to infrequent or limited services. SAVs could supplement existing public

56 REPORT transportation services by providing additional options in less dense communities. In some areas, a fleet of SAVs may be more affordable and better suited to the population density than maintaining fixed-route transit services. By removing the human driver, SAVs may also be a more affordable option to fill service gaps and provide riders with on-demand options if fixed-route options are not available. In September 2019, the USDOT funded two research projects with DriveOhio and Texas A&M University that focus on AV technology in rural settings (Descant, 2019). 5. Off-peak/temporal gap-filling services: Similar to spatial gap-filling services, SAVs could provide an additional, on-demand or fixed-route service to those traveling during off-peak hours. SAVs may be able to better meet different rider demands during these times by providing on-demand rides in the areas needed. In addition, by removing the human driver, SAVs may be a more affordable option (for public agencies to provide and riders to own) to fill service gaps. For example, Transdev (an international mobility company) is piloting a SAV shuttle service in the Paris-Saclay area of France between the hours of 12:30 AM and 3:00 AM. The goal of the pilot is to provide an additional transportation option outside the hours of operation for the existing transportation services (Transdev, n.d.). Providing New Services In addition to supplementing public transit, SAVs may also be able to offer travelers new services. These services may enhance traveler mobility and the accessibility of destinations and resources. 1. Closed campus: SAVs could provide short-distance, point-to-point travel in closed- campus environments. These locations include theme parks, resorts, malls, business parks, college campuses, airport terminals, construction sites, downtown centers, real estate developments, gated communities, industrial centers, and others. In February 2019, COAST Autonomous piloted an AV that could hold eight people on the University of South Florida’s campus. The pilot lasted a week, and the AV operated in an area with pedestrians, bicyclists, and other low-speed transportation modes (Nghiem, 2019). 2. Emergency response: In the future, SAVs may provide emergency transportation. One potential use case is an automated ambulance that determines the most efficient path from a crash scene to a treatment center. In 2018, Ford submitted a patent for an automated police vehicle and a patent for routing technology that could be used for ambulances and fire trucks. 3. Urban goods delivery: CNS, also referred to as flexible goods delivery or app-based deliveries) provide for-hire delivery services for monetary compensation using an online application or platform (such as a website or smartphone app). A CNS connects couriers (who typically use their personal vehicles, bicycles, or scooters) with freight (e.g., packages, food). In the future, these services could be paired with ADVs and robots. For example, Nuro partnered with Kroger’s Fry Foods grocery store in Scottsdale, Arizona, in September 2018 for a six-month pilot program to offer automated grocery deliveries. The pilot first used automated Toyota Priuses to deliver groceries to customers, but it later shifted to Nuro’s exclusively designed R1 model (Figure 15) (Nuro, 2018). In March 2019, customers who participated in the Scottsdale pilot switched over to an existing grocery delivery service provided by Kroger in partnership with Instacart (Wiles,

57 REPORT 2019). Another example of a CNS is Kiwibot. Since its start in 2017, Kiwibot made over 30,000 deliveries, built over 150 robots, and serves college communities around the country (Kiwibot, n.d.). During the COVID-19 pandemic, JTA used AVs to deliver medical supplies and tests for the COVID-19 virus to a local private hospital. In addition to NAVYA, JTA began working with Beep, an AV fleet provider, to use AVs to drive COVID-19 tests from testing centers to the local hospital (Scanlan, 2020). Figure 15. Nuro Delivery Vehicle Source: Nuro, 2018 Partnerships Partnerships can enable use cases for MOD and AVs. These emerging transportation modes face an array of policy, financial, acceptance, and marketing challenges. These challenges may be addressed through partnerships that can assist with risk sharing, financial subsidies, and joint marketing. Additionally, partnerships can lay the foundation for establishing standards, data sharing, integrating new modes in the public rights-of-way, and supporting the development of public policies. Partnerships for MOD and SAVs can engage a variety of stakeholders with different responsibilities. Partners may regulate pilot programs, facilitate data sharing, provide or connect to public transit, offer information and fare payment services, and assist with marketing. Table 12 lists common partners, types of support, and examples.

58 REPORT Table 12. MOD and SAV Partnerships Partners Types of Support Support Example(s) C ity a nd R eg io na l G ov er nm en ts a nd P ub lic A ge nc ie s Business Health In 2020, the San Francisco Municipal Transportation Agency began closing 2.2 miles of Market Street in San Francisco to private vehicles to support bicyclists and pedestrian safety and business health (SFMTA, 2019). Data Sharing The City of Los Angeles shares rights-of-way capacity and incident data with MOD apps. Managing Rights-of- Way Several communities have developed permitting processes for shared micromobility operators to complete before entering the local market (e.g., San Jose, Seattle). Inclusion of MOD/SAVs as a Transportation Demand Management (TDM) Measure Santa Monica, California requires developments seeking a variance to include TDM measures. TDM measures can include incorporating shared modes, carpooling parking, bicycle lockers, or workplace showers (to encourage cycling). Facilitating Stakeholder and Community Involvement The New York City DOT conducted over 100 public meetings in multiple languages along with virtual engagement as part of the city’s bikesharing planning process. Risk Sharing Ford Motor Company partnered with the Miami-Dade County government and Domino’s Pizza to test delivering pizza in AVs in Miami, Florida (Associated Press, 2018). The company engaged in a partnership with the city to attempt to find ways to address congestion, as well as to test emerging AV delivery services for viability and profitability. Pu bl ic T ra ns it O pe ra to rs Data Sharing The regional transportation district (RTD) in Denver shares real-time public transit information vis-à-vis a web-based trip planner. In early 2019, RTD began sharing real-time data with Uber in a partnership to allow riders to plan trips and buy tickets within the Uber app (Bosselman, 2019). MOD Linking In 2019, DART signed a contract with Uber, which offers free and discounted shared rides in some regions of the Dallas area. This new contract also makes Uber rides originating or terminating at DART stations free (Repko, 2019). Transit Fare Integration In June 2020, New York City’s Metropolitan Transportation Authority (MTA) launched OMNY, an integrated contactless fare payment. Travelers can use credit, debit, and prepaid cards; digital wallets; mobile phones; smart watches; and other wearable devices to pay for fares for any of MTA’s subways, buses, or commuter railroads. Designated Parking or Loading Space BART provides designated carpooling spaces at select BART station parking lots. MOD Fare Subsidy As part of the MOD Sandbox project, PSTA provides subsidies to paratransit and TNC operators when providing service to low- income users and people with disabilities. Paratransit Many public transit agencies subcontract to third-party vendors to provide paratransit services (e.g., PSTA).

59 REPORT E m pl oy er s a nd B us in es se s Internal Marketing The Seattle Times provides internal marketing support for carsharing. Use of MOD for Business Travel Swedish Medical Center in Seattle provides carsharing memberships to employees for business-related trips. Incorporating MOD into Facility Design In Downtown Berkeley, California new apartment buildings that are equipped with parking are required to reserve parking spaces for carsharing services. MOD Membership Subsidies Condominium developers in Vancouver, Canada, offer carsharing membership as an amenity to residents. Risk Sharing Developers can employ risk-sharing partnerships where the cost of MOD service is subtracted from monthly revenue and billing any shortfall to the risk partner (e.g., developer/property manager). Information Displays TransitScreen, a software company that provides a digital display of real-time transportation information, has displays in over 1,000 buildings in 30 cities in the US (TransitScreen, 2020). Marketing Advertisements at Philadelphia’s bikesharing program Indego’s docking stations help the service generate revenue. U ni ve rs iti es Research Partnerships NAVYA provided automated shuttles for services on the University of Michigan campus; this pilot ended in December 2019. The pilot project provided important research and data collection on user behavior (University of Michigan, 2019). Marketing to Students The micromobility company Superpedestrian is partnering with the carsharing company Zipcar to offer bikesharing services to 15 universities and communities in the US (Beyer, 2017; O’Kane, 2020). Designated Parking George Mason University designated bike racks and corrals for shared micromobility devices (e.g., bicycles, scooters) to be parked. Supporting SAVs through partnerships may result in social, transportation network, and other benefits including: Individuals • Increased modal and multimodal options; • Congestion mitigation; and • Cost savings. Transportation Network • Decreased need for parking, • Cost savings, • Bridged gaps in the transportation network (e.g., first- and last-mile connections), • Increased efficiency, • Fare integration with other modes, and • Increased public transit ridership. Organizations • Increased economic activity near SAV services and modal nodes; • Improved efficiency of existing fleets (e.g., motor pools); Partners Types of Support Support Example(s)

60 REPORT • Opportunities for discounts, joint marketing, education, and outreach; • Tax savings for property managers and employers; • Greater number of supporters of shared modes (e.g., employers, institutions); and • Reinforced image of sustainability and corporate stewardship. Environment • Improved air quality, and • Reduced GHG emissions. Social Equity • Improved accessibility and mobility; and • Increased vehicle and goods access, particularly for carless households. Agencies, such as the SFCTA, are engaging in partnerships in an attempt to gain these benefits. SFCTA is planning to run an automated shuttle on the nearby Treasure Island to reduce traffic congestion and enhance connectivity to ferry and bus service in San Francisco and the East Bay. The shuttle is funded by a $1 million grant from the USDOT in addition to local (both public and private) matching funds and is anticipated to begin operations in 2021 (Siu, 2017; SFCTA, 2019). While most partnerships between public agencies and MOD and SAV providers are informal (e.g., joint marketing), public agencies are increasingly entering more formalized partnerships through requests for proposals (RFPs), memoranda of understanding, and other processes. For example, in 2018, the JTA issued a procurement notice to modernize its aging Skyway monorail service with SAVs. In early 2018, Transdev operated a six-month automated shuttle pilot in Jacksonville as preparation for modernizing the Skyway for SAVs (Robinson, 2018). Subsequently, in December 2018, JTA announced the testing of a second automated shuttle model that is manufactured by NAVYA. Currently, the test vehicle is running along a test track under local expressway ramps. JTA expects the first leg of the automated system to be open before the end of 2020. JTA and the city recently secured $25 million in federal funding to help fund the automated, EV mass transit project (Bortzfield, 2019). Built Environment AVs may be deployed in a variety of business models and use cases in different built environment types with different infrastructure considerations. As depicted in Figure 16, over the past few decades, the United States has witnessed the development of a variety of built environment types including city center, suburban, edge city, exurban, and rural.

61 REPORT Figure 16. Five Common Built Environments Source: Shaheen and Cohen, 2016; NCHRP, 2020 These built environment types are associated with their own unique challenges. Table 13 summarizes the built environment types and associated challenges. Table 13. Common Built Environment Types and Challenges Built Environment Type Definition Challenges City Center A development framework with the highest concentration of jobs, comprised of Central Business Districts and surrounding neighborhoods. Limited parking and loading zone capacity and peak hour roadway congestion and transit congestion Suburban A less urbanized development pattern with high levels of low-density residential uses with fewer jobs than residences. Limited or infrequent public transit service and a built environment that is more conducive to privately owned vehicles Edge City An urbanization pattern presenting some features of city center employment mixed with suburban form. Edge cities are often built around highway interchanges (and occasionally around rail stations) with higher concentrations of office and retail space often paired with multi-family residences. High congestion and a built environment not generally conducive for active transportation Exurban A low-density residential development within the commute shed (area) of a larger and denser urbanized area. Long commute distances and limited public transportation

62 REPORT Built Environment Type Definition Challenges Rural The lowest density development pattern characterized by low-density light industrial, agricultural, and other resource-based employment. Long travel times between jobs, healthcare, and retail centers with limited public transportation options often necessitating private vehicle ownership Source: Shaheen et al., Forthcoming As part of the post-war growth of suburbs, edge cities, and exurbs, the United States has emphasized auto oriented infrastructure including highways, motels, drive-thrus, gas stations, and parking structures. In an automated future, some people may move closer to city centers as AVs create new opportunities to replace urban parking with housing and other land uses. However, some people may move farther from urban centers into suburbs and exurbs if long commutes become more palatable when travelers no longer need to drive themselves. Growth patterns may also change with automated goods delivery which may allow people to live further from resources, such as grocery stores. AVs could impact, or be impacted by, each built environment type through deployment purposes including: • City center markets: Zero-, single-, or low-occupancy AVs have the potential to exacerbate the high density of trip origins and destinations in the city center built environment (Shaheen et al., 2017b). Some worry that a shift in travel behavior (e.g., travelers moving from high-occupancy public transit to low-occupancy vehicles) and in commuter behavior (e.g., travelers commuting from farther reaches into urban cities due to the convenience of AVs) could have negative repercussions for congestion and environmental impacts. However, AVs could offer potential benefits to cities and travelers. SAVs could encourage individuals to forego private vehicle ownership, reducing the need for parking and encouraging the use of shared modes. SAVs may also increase the affordability of transportation (by providing additional modal options at a variety of price points), housing (through reduced parking infrastructure, if savings are passed on to consumers), and eliminating trips (through goods and service delivery options). • Suburban markets: In this environment, AVs can help those unable to drive or those without vehicles. However, in a suburban market, SAVs may be less affordable due to factors including longer trip lengths and a smaller customer base. Some worry that the convenience of AVs will encourage individuals to live farther from work and other amenities, increasing VMT. • Edge city markets: Potential benefits of SAVs in edge cities include increased modal alternatives and options for those who cannot drive, reduced travel times (e.g., travelers do not need to find parking for their vehicle), and reduced vehicle ownership. However, if vehicle automation increases vehicle use, increased congestion and conflict at the curb could occur, especially during commute hours. Like suburbs, SAVs may be less affordable due to low population density and longer travel distances. • Exurban markets: Exurban SAVs can enhance mobility for those without vehicles by providing lower-cost and more frequent or responsive transit strategies. • Rural markets: Compared with urban and suburban areas, research and pilots focusing on AV applications in exurban and rural communities are limited. More research and

63 REPORT pilots are needed to test AV applications and use cases that target these built environments. For example, a rural SAV pilot may require a flexible fare structure to serve low-income households. Operational characteristics, such as willingness to wait for an SAV in rural areas, may differ from traveler expectations in urban areas. Additionally, rural SAV applications may focus on social connections, reducing isolation, and goods and service delivery, rather than reducing pick-up and travel times (Berger, 2018). Increasing mobility has the potential to expand rural transit, goods delivery, and other services that can increase economic activity and quality of life for rural communities (Berger, 2018). However, the introduction of AVs has the potential to increase reliance on personal vehicle trips if AVs become widely affordable. AV Interaction with Infrastructure Regardless of built environment type, infrastructure owner operators (IOOs) will be an important factor in understanding how AVs interact with infrastructure. IOOs invest in, develop, service, maintain, and operate infrastructure. IOOs can play a role in advancing, operating, and maintaining the physical and digital infrastructure for MOD and AVs. There are opportunities for IOOs to make infrastructure investments to improve the safety and efficiency of AVs including maintaining markings and signs, dedicating lanes, and providing data feeds for work zone information. AVs can also provide benefits to IOOs by providing probe data of infrastructure conditions, identifying hazards, and informing emergency response. Automakers and technology companies are currently developing AVs that are primarily electric. This results in the need for the development and implementation of electric infrastructure elements. AVs may require electric charging stations, strong electricity grids, and backup generators or infrastructure elements in case of emergency (e.g., electricity blackouts). Also, AVs will likely require electric infrastructure that allows them to charge quickly. These considerations may impact infrastructure development and AV travel patterns. For example, if electric charging stations are not easily available in rural areas or along long stretches of highway, long-distance trips may not be possible in AVs. In addition, AVs will need to be storage locations (e.g., parking lots, garages) when not in use to ensure they are not roaming and creating congestion. These storage areas may double as charging stations. AVs that are also connected vehicles will need supportive infrastructure elements including roadway sensors, internet connections, and databases to store and process communication data. These infrastructure elements will also need to be maintained through the deployment and continued use of AVs. Infrastructure elements that are necessary to support AV operations may require public agency support. Agencies may need to assist in financing development, advocating for their development, and leveraging policies and/or incentives to encourage private developers to include these infrastructure elements in new and redeveloped areas. Financial support may be generated through the restructuring of internal budgets, engaging in partnerships, and altering the revenue-generating process. Section 3 of the Shared Automated Vehicle Toolkit provides further information on potential built environment and infrastructure considerations as well as potential strategies to fund these changes.

64 REPORT Policies and Regulations In addition to business models, use cases, the built environment, and infrastructure considerations, agencies at the local, regional, and state level can have a notable impact on the success of AV operations through public policy, legislation, and regulation. Public agencies may regulate AV operations to protect the health and safety of both travelers and non-travelers (e.g., bystanders). Often, it may not be clear who is responsible for regulating AV operations since regulatory oversight could be the responsibility of numerous overlapping levels of government (e.g., local, state, and federal) and agencies or necessitate the development of a new regulatory framework for an automated future. Another potential approach is deregulation or self- regulation, such as the enforcement of the US Federal Motor Carrier Safety Standards. Additionally, public agencies will need to identify what is being regulated. Regulations could encompass a variety of operational areas, such as the vehicles, fares, occupancy, or SAV operators’ business practices. For example, regulations may need to be developed to support multi-occupancy trips to avoid the potential challenges resulting from zero- and single- occupancy trips (e.g., roadway congestion, unavailability of vehicles). Regulations that specifically govern the development of AVs may need to mandate that they are designed to be fail safe10 rather than fail fatal.11 As AV use cases expand to include services, such as goods delivery, federal and state policies will likely need to regulate and guide these developments as well. Existing AV legislation in the United States typically focuses on safety, liability, insurance, vehicle design, and operations. Many of these policies have been enacted at the state or local level. At the federal level, AV legislation passed the House of Representatives but stalled in the Senate in early 2018. In late 2019, Congress started a new bipartisan initiative to draft AV legislation. In the meantime, there have been attempts by federal agencies to create a framework for AV policy. In October 2018, the USDOT released ADS 3.0, which builds upon the USDOT’s 2.0: A Vision for Safety and provides guidance to states on the training and licensing of test drivers. ADS 3.0 also offers guidance on driver engagement methods during AV testing (NHTSA, 2018). In January 2020, the USDOT released Ensuring American Leadership in Automated Vehicle Technologies: Autonomous Vehicles 4.0 (i.e., AV 4.0). The report defines the three principles the US government is adhering to during the development of AVs: 1) protect users and communities, 2) promote efficient markets, and 3) facilitate coordinated efforts (National Science and Technology Council and USDOT, 2020). Federal agencies may be able to refer to other industries that have also undergone automation and/or centralization (e.g., aviation policies) for guidance on vehicle automation programs and policies. Table 14 lists federal AV policy actions. 10 A design feature that ensures that if a device fails it will respond in a way that will cause minimal or no harm to the user. 11 If a design is not fail safe it may fail in a way that results in injury or damage to the user.

65 REPORT Table 14. Federal AV Regulation, Guidance, and Policy Title Agency/ Government Entity Status Description Self-Drive Act (2017) House of Representatives Passed House, failed Senate This would establish a federal framework for AV regulation and dramatically increase exemptions for AVs from the Federal Motor Vehicle Safety Standards (Stocker and Shaheen, 2018). AV Start Act (2017-18) Senate Stalled This act would allow operators to test and market AVs immediately before federal rulemaking is established. However, it was stalled due to safety concerns (Bonazzo, 2018). Federal Automated Vehicles Policy (2016) USDOT Adopted This established a voluntary 15-point framework with recommended guidance for future policy (Stocker and Shaheen, 2018). Automated Driving Systems 2.0 (2017) USDOT Published This established a voluntary 12-point framework and clarified that there was no need to wait for federal approval to test or deploy AVs. Many companies have submitted the Voluntary Safety Self-Assessment letters from the framework including Waymo, GM, and Ford (USDOT, 2017). Data for Automated Vehicle Integration (2017-present) USDOT Ongoing Initiative In December 2017, the USDOT hosted the Roundtable on Data for Automated Vehicle Safety. This was followed by a Public Listening Summit on AV Policy to inform the development of ADS 3.0 in March 2018. The purposes of these efforts were to: 1) solicit feedback on the USDOT’s draft Guiding Principles on Voluntary Data Exchanges to Accelerate Safe Deployment of Automated Vehicles and draft Framework for Voluntary Data Exchanges to Accelerate Safe Deployment of Automated Vehicles; and 2) identify near-term data exchange use cases to accelerate the safe rollout of AVs. The USDOT recently announced plans to fund Work Zone Data Exchange (WZDx) Demonstration Grants to provide one-time funding for public roadway operators to make unified data feeds available for third-party use. Impact of Automated Vehicle Technologies on Workfoce USDOT Ongoing Initiative In 2018, $1.5 million was allocated to the USDOT, in partnership with other agencies, for a comprehensive analysis on the impact of Advanced Driver Assist Systems (ADAS) and HAV technologies. The study is focuses on four major areas: 1) labor force transformation/displacement, 2) labor force

66 REPORT Title Agency/ Government Entity Status Description (2018 – present) training needs, 3) technology operational safety issues, and 4) quality of life effects due to automation. The first phase of the study focuses on trucking and transit bus sectors, while the second phase includes broader driving occupations and potential impacts to supportive industries. Inclusive Design Challenge (2020 – present) USDOT Ongoing Initiative The Inclusive Design Challenge is focused on identifying innovative design strategies that can enable people with a variety of disabilities to use AVs to access employment opportunities, healthcare centers, and other critical destinations. The goal of the challenge is to address barriers including locating, entering, and exiting AVs; securing passengers and their associated devices; entering information; and interacting with AVs in a variety of environments. Preparing for the Future of Transportation: Automated Vehicles 3.0 USDOT Published AV 3.0 expands the scope of AV 2.0 to all surface on-road transportation systems. It was developed through input from a diverse set of stakeholder engagements throughout the US, AV 3.0 aims to focus on three key areas: advancing multimodal safety, reducing policy uncertainty, and outlining a process for working with USDOT (USDOT, 2018). Ensuring American Leadership in Automated Vehicle Technologies: Autonomous Vehicles 4.0 USDOT Published AV 4.0 builds off of existing publications and expands the scope of AV 3.0 and AV 3.0. The goal of the report is to address safety concerns associated with the development and deployment of AVs. AV 4.0 states that the USDOT is working on establishing performance, manufacturing, and operational standards in order to increase safety in AV testing and regulation (National Science and Technology Council and USDOT, 2020). Bipartisan Autonomous Vehicle Bill House of Representatives and Senate House and Senate Committee drafted legislative text Congressional committees have been engaged in a bipartisan and bicameral initiative to draft AV legislation that will create a path to deployment to ensure companies continue to develop AV technologies in the US as of October 2019, Congress has solicited input from stakeholders and circulated draft text for three sections of the bill. More sections are expected to follow (Rogers, 2019).

67 REPORT In the future, the federal role regarding AVs could evolve to encompass new responsibilities. In addition to developing regulations, federal agencies may take other actions including: • Developing universal design12 standards: The emergence of innovative transportation modes offers federal agencies the opportunity to define design standards that help ensure that these modes are accessible by a variety of users. • Facilitating peer exchange: As AVs continue to rapidly develop, federal agencies may be able to facilitate and support peer exchange to ensure that case studies, lessons learned, and best practices are disseminated to stakeholders. • Granting permits: Agencies at the federal level may be able to develop permits and permitting processes to allow the testing and operations of AVs in different areas including public roads. • Leveraging taxes: Agencies may levy taxes on AV operators and riders to finance the development of supportive infrastructure for these modes and address revenue losses (e.g., repurposing current curbside parking for passenger loading zones). • Providing funding: Federal agencies may be able to offer supplemental funding to state, regional, and local governments to ensure that AV fleets can serve a variety of communities (e.g., people with disabilities, rural communities, low-income neighborhoods). • Supporting rural programs: AV pilot programs and projects are predominantly taking place in higher density, urban environments. However, there are a variety of ways these transportation modes could benefit less dense, rural areas. As a result, federal agencies can support pilots taking place in rural environments through actions, such as issuing RFPs, allocating funding, and disseminating best practices. For example, one of the USDOT Automated Driving System Demonstration grants focuses on AV technology in rural settings. In addition to federal policies, state agencies are also beginning to develop AV-related policies. All but nine states have enacted state policies related to AVs. Thirty states and the District of Columbia have enacted state legislation, six states have had their governor pass an executive order pertaining to AVs, and five states have both state legislation and executive orders. The map in Figure 17 shows legislation and executive orders enacted in different states as well as whether these laws/orders require an in-vehicle operator to test AVs. The figure also distinguishes states that have legislation pertaining only to vehicle platooning or commercial vehicle operation. Since 2017, 45 states introduced AV-related legislation (National Conference of State Legislatures, 2020). In 2018, 15 states enacted AV-related bills, and as of August 2020, 22 states have enacted AV-related bills (NCSL, 2019; National Conference of State Legislatures, 2020). 12 Universal design is a design principle focused on designing an environment that is accessible and usable by people with a variety of needs and capabilities.

68 REPORT Figure 17. Map of State-Level AV Legislation and Executive Orders as of December 2019 States have addressed the oversight of on-road testing of AVs by companies in different ways. Some states, such as California, have implemented strict regulations regarding the registration and operation of AVs. Others including Florida and Arizona have taken a different approach with limited or no state legislation. California has implemented a permitting process that requires companies to register the vehicle and track performance issues, such as disengagements (California DMV, 2018). Restrictions on AV testing (e.g., the presence of a steering wheel or test driver) were loosened through subsequent legislation. In 2018, California passed two bills related to enforcement of AV permits and a tax on trips taken in AVs. Most recently, the California DMV announced a permitting process for automated light-duty commercial trucks. In contrast, Arizona has no state legislation on AVs and has acted through executive orders. Arizona’s governor signed an executive order in 2015 that directed state agencies to undertake steps to support the testing and operation of AVs and enabled pilot programs. In 2018, Arizona’s governor issued two executive orders: one requires AVs to comply with federal and state safety standards, while the second establishes a research and testing institute. Many other states fall along the spectrum of these two examples. In addition, several states including Alabama, Kentucky, Louisiana, Minnesota, Mississippi, Oklahoma, and Wisconsin, only address AVs through legislation on truck platooning or commercial services. Only a few states have enacted legislation regulating SAV fleets. Nevada and Tennessee both passed taxes on SAV services. Nevada’s Assembly Bill 69, which passed and was approved in 2017, imposes a three percent tax on single passenger SAV fares and requires vehicles to accommodate wheelchair accessibility. In April 2016, Tennessee passed Senate Bill 1561, which establishes a one cent per-mile tax for AV passengers and a 2.6 cent per-mile tax for automated trucks. The revenue from this fund is shared between the general fund, the transportation department, and local governments (Green, 2018). In California, now-retired Governor Jerry Brown signed Assembly Bill 1184 (AB 1184) in September 2018 (Ting, 2018). AB 1184 authorizes

69 REPORT the City and County of San Francisco to tax AV rides originating in San Francisco, whether facilitated by a TNC or another person. Under the bill, the AV tax is capped at 1.5 percent of net rider fares when a passenger shares a ride, and 3.25 percent of net rider fees when the passenger does not share the ride. AB 1184 requires San Francisco to use the revenues from this tax to fund transportation operations and infrastructure within the city (Ebbink, 2018). In contrast, Alabama, Florida, and Nebraska have passed legislation prohibiting local governments from imposing taxes on AVs (though the bills do not exempt AVs from taxes applied to non-AVs) (NCSL, 2019). As more companies test AVs in urban settings, regional and state agencies may need to address conflicting policies and develop a unified regulatory framework. A unified regulatory framework and predetermined standards can help create consistency across a variety of cities, counties, and states. In Boston, the Massachusetts DOT still plays a crucial role in approving AV testing applications and assisting local agencies to develop their own AV testing phase plans. While local municipalities and states tend to work together, this may not always be the case. For example, Texas allowed Uber and Lyft (non-AVs) to operate in Austin, overruling a local regulation (Sisson, 2017). This demonstrates states’ ability to pre-empt local ordinances, either allowing or disallowing AV testing in contravention of local policies. User age is another legal consideration for chartering SAV services. Uber’s and Lyft’s Terms of Use have age restrictions. For example, Uber states, “you must be at least 18 years of age, or the age of legal majority in your jurisdiction (if different than 18), to obtain an account, unless a specific service permits otherwise” (Uber, 2018). The question remains whether SAV services will continue to use this policy and if actionable guidance and recommendations for parents and guardians to chaperone minor children will be needed. Standard Terms and Definitions In addition to guiding policies, standards represent a key enabler of MOD and AVs. SAE International has established standards for the six levels of AV automation. In October 2018, SAE International announced new definitions for shared mobility terms in the Taxonomy and Definitions for Terms Related to Shared Mobility and Enabling Technologies— J3163™ standard. The J3163™ standard defines shared mobility as “the shared use of a vehicle, motorcycle, scooter, bicycle, or other travel mode; it provides users with short-term access to a transportation mode on an as-needed basis.” The J3163™ standard includes definitions for shared modes (e.g., carsharing, bikesharing) and enabling technologies. Rights-of-Way Regulation Policies and standards may need to regulate the impact of AVs on the rights-of-way. MOD, AVs, and automated goods delivery services can have several impacts on rights-of-way (e.g., curbspace, loading zones, and parking). Potential impacts include: 1) increased use of the rights- of-way that create competition among modes and service providers for a limited amount of space; increased modal and operator activity resulting in safety hazards, such as modal conflicts and congestion in high-traffic locations; and unintended impacts on vulnerable communities including MOD or AVs blocking ADA access to the rights-of-way. These impacts are encouraging communities to reconsider rights-of-way regulations that have typically been

70 REPORT managed based on adjacent land use and historical precedent. Often, regulations on curbspace use are static, leaving communities unable to adjust prices based on the time of day or on- demand (NACTO, 2019). The following sections describe various policies for rights-of-way allocation, pricing for rights-of-way access, methods for managing competition for curbspace, and recommendations for rights-of-way management in an AV future. Policies for Rights-of-Way Allocation To address the challenge of allocating limited curbspace, public agencies have developed formal and informal policies to distribute rights-of-way to different modes. Table 15 displays various considerations for communities to consider when allocating the rights-of-way. Table 15. Considerations in Rights-of-Way Allocation Considerations Description Service Characteristics (e.g., hourly rentals, membership-based services) • Business models (e.g., for profit, nonprofit) Procedures for Allocating Rights-of-Way • Jurisdiction (e.g., city staff, city council, parking authority) • Process (e.g., first-come, first-serve; lottery; auction; request for proposal or pilot) Methodology for Valuing Rights-of-Way • Cost recovery of program administration • Foregone meter, permit, and other revenue • Supply and demand (e.g., auctions) • For profit (e.g., generate revenue for local coffers) Performance Metrics • Measuring system performance through metrics • Performance metrics can include person throughput (i.e., number of people or vehicles per hour) and modal goals (i.e., number of bikesharing devices used per hour) Management of Competition • Methods for managing competition between operators • Methods for managing competition between modes • Method for dispute resolution (e.g., administration hearings/appeals, mediation, arbitration, litigation) Source: Shaheen et al., Forthcoming The rights-of-way can be managed through a variety of policy approaches including: • Developing a process for accessing and using the public rights-of-way; • Identifying permits that can be issued or fees that may be charged for mobility service providers operating in the public rights-of-way; • Establishing standards for signage and/or markings to identify proper parking areas for vehicles and devices (e.g., bicycles and scooters); • Enforcing loading and parking compliance through virtual geographic boundaries (commonly referred to as geofencing) using GPS, radio frequency identification, or other technologies; and • Employing data sharing requirements and/or requiring impact studies as a condition for allowing services to use the public rights-of-way.

71 REPORT Even with these management strategies, communities will likely encounter competing demand for the curbspace. To mitigate conflicts at the curb and provide fair access opportunities, communities will need to develop practices for managing competition. Pricing for Rights-of-Way Access Rights-of-way access may be managed through a variety of pricing approaches that focus on congestion, trip, and incentive pricing approaches. Congestion pricing addresses congestion challenges by pricing zones, rights-of-way, loading zone use, and/or performance. Trip pricing focuses on trip types and distances to encourage the use of select modes, services (e.g., ridesharing and vehicle sharing), and trip distances. Incentives are a pricing tool that can be used to support certain modes, times of operations, trip types, or areas of travel. Incentives often include discounts or money back for passengers and drivers. These tools can address congestion in high demand areas (e.g., downtowns) or during high demand times (e.g., commute hours). Section 3 of the Shared Automated Vehicle Toolkit contains a further explanation and example of rights-of-way pricing. Managing Competition for the Curbspace When allocating the increasingly crowded rights-of-way, communities may also need strategies to manage competition among various transportation modes and providers. Automated goods delivery services may create additional competition for curbspace. A variety of methods for managing competition are described below (Cohen and Shaheen, 2016): • First-come, first-serve: A public policy where requests for public rights-of-way by private operators are attended to in the order in which they arrive; • Lotteries: A public policy where requests for rights-of-way are selected by random drawing; • Auctions: A public policy where requests for rights-of-way are granted to the highest bidder; • Preferential treatment: A public policy that gives preferential treatment to a specific mobility operator for a reason; • Collaborative approaches: A public policy employing a collaborative process, such as negotiation or mediation, to reach a mutually beneficial outcome; • RFPs: A solicitation, often through a bidding process, by a public agency or government interested in procuring a mobility service; and • Tandem policies: A public policy where every stakeholder receives an equal share of the public rights-of-way. Rights-of-Way Management in an AV Future The future impacts of AVs on the public right-of-way are uncertain. Because AVs do not need to park at a traveler’s destination, vehicle automation may allow for more efficient curbspace, parking, and fleet management use. For example, idle vehicles may be able to park themselves in less congested areas, rather than circling to find a spot at a crowded destination (International Transport Forum, 2018). In addition, AVs can help communities catalog and manage the rights- of-way by sharing data collected during testing and road mapping (see the text box below for

72 REPORT more details on AVs for data collection). However, AVs may also have unintended impacts on the rights-of-way. Some researchers predict that AVs may encourage the use of single- occupancy vehicles and create congestion in the right-of-way. Communities can attempt to mitigate and prevent negative impacts through informed policy decisions that consider current curbspace use and planning goals (e.g., reduce vehicle emissions, increase accessibility). Some experts recommend prioritizing curbspace as a potential management strategy, especially as evolving MOD and SAV services encourage communities to reconsider traditional designs and emphasize pedestrian and bicycle access. The National Association of Transportation Officials’ Blueprint for Autonomous Urbanism: Second Edition (2019) recommends curbspace prioritization in the following order of importance: 1. Buses, transit, and bicycles should receive priority as sustainable and efficient modes. 2. Freight and delivery should next be given priority, as they enable economic activity. NACTO also suggests providing green spaces for social activities and shopping. 3. Individual trip vehicles; inclusive of personal, shared, fleet, and TNC trips; should be given the last priority due to the inefficiency in the number of people served. While NACTO does not distinguish between individual trip vehicles, a further recommendation is to prioritize shared or pooled vehicle access to the curb over personal vehicle access. Many additional strategies exist for managing the public rights-of-way including but not limited to the approaches listed previously: assigning access on a first-come, first-serve basis; holding lotteries or auctions; employing a collaborative process, or soliciting operators through an RFP.

73 RT Communities can select a policy approach based on shared goals, service characteristics, and current or planned allocation procedures. Privacy and Data Sharing Policies and regulations may also need to address AV data sharing. Throughout testing and operations, AVs collect large quantities of data on road infrastructure and routes from vehicle sensors, lidar, and cameras. In addition, AV services collect data on users—including trip routes, pick-up/drop-off locations, speed, and cost—mainly through GPS data from mobile phones and vehicles. This data is useful for AV providers, allowing them to optimize operations and improve AV performance and safety. Similarly, members of the public and academic sector will likely want access to data to help manage supporting infrastructure, ensure that these services are meeting service requirements (e.g., meeting equity or environmental goals), and understand AV impacts. In the future, policymakers may need to consider policies regarding data and AV services, as data collection and sharing may pose privacy challenges (discussed in further detail in later sections). The following subsections discuss challenges for data collection and sharing and propose strategies for protecting user privacy. Privacy Challenges for Data Collection and Sharing There are three main concerns with data collection and sharing for SAV services. First, companies may collect PII with or without a user’s knowledge. Secondly, trip data can become PII through routine travel patterns (e.g., detailed GPS data of users traveling from home to work, school, stores, or other locations) or when combined with other data (e.g., addresses, phone numbers, demographic data) (NACTO, 2019). Lastly, private companies may be wary of sharing data with the public sector due to concerns over sensitive data. Mobility and goods service providers may generate or rely on proprietary information (e.g., business strategies, trade secrets, routing algorithms). This proprietary data may be important to a company’s business plan, raising concerns over what type of data is shared. These companies may want a certain level of aggregation to avoid exposing sensitive material through public records. Public record laws that allow people to submit public record requests may challenge public agencies who are working to protect traveler privacy or company trade secrets. Under the federal Freedom of Information Act and many state public records laws, operators that receive public funding or partner with public agencies may have their data become a public record and be subject to public release (Kimley Horn and IBI Group, 2019). The following section details policy recommendations to address these privacy challenges. Potential Policies Several nonprofit organizations, public agencies, and companies have released recommendations for data management including: • Access: Develop policies to ensure access to short- and long-term communication systems, create universal agreements on data availability; • Aggregate: Aggregate data to protect individual information; • Audits: Use outside organizations (e.g., nonprofits, other jurisdictions) to conduct audits for security, accuracy, and efficacy;

74 REPORT • Capacity: Increase internal capacity for data management; • Categorization: Categorize different types of data and specify their uses; • Consequences: Enforce consequences (e.g., revoking permits) for organizations who do not comply with the predetermined data sharing agreements; • Designation: Designate trip data as confidential information; • Feedback: Offer opportunities for organizations and customers to provide feedback; • Health and Safety: Design and implement protocols that protect the wellbeing of drivers, operators, and passengers; • Privacy: Develop privacy protection policies; • Security: Implement appropriate safeguards to ensure data security, require sensitive data to only be shared through secure data feeds, establish licensing and security provisions; • Sharing: Only require the sharing of data for specific operational needs; • Standards: Ensure data adheres to standards, is open whenever possible, and protects travelers privacy and data; • Storage: Use third parties to store data; and • Transparency: Publish a list of data collection types and time frame of collection. Source: LADOT, 2019; NACTO, 2019; Clewlow, 2019; Transportation for America, 2019, Bailey, 2018 While there are a limited number of case studies of communities implementing the above practices, some communities mandate data sharing as a requirement for AV testing. For example, Boston, Massachusetts requires that companies testing AVs in the city provide quarterly data reports that are publicly available. These reports are required to include information on miles driven, locations driven to, crash reports, and situations where the engineer had to take over for the automated system (nuTonomy, 2018). The public reports are short five-page summaries; however, the Massachusetts DOT receives additional information including the number of passenger trips, passenger home zip codes, trip origin and destination, and qualitative user feedback (Fiandaca, 2017; Fiandaca, 2018). Concerns over user data privacy have led to important developments in privacy laws that could impact open data sharing. The European Union’s (EU’s) General Data Protection Regulation (GDPR), which became effective in May 2018, establishes a set of standards for how companies handle EU citizens’ data to protect consumer and personal data. Key requirements of GDPR include: requiring the consent of subjects for data processing, anonymizing collected data to protect privacy, providing data breach notifications, safely handling the transfer of data across borders, and requiring certain companies to appoint a data protection officer to oversee compliance. GDPR applies to any company that markets goods or services to EU residents, regardless of its base location. As a result, GDPR has an impact on data protection requirements globally (De Groot, 2019). In 2020, California will implement Assembly Bill (AB) 375, or the California Consumer Privacy Act, which sets forward similar requirements to GDPR. AB 375 classifies geolocation data as personal information. In addition, consumers can request to know what data is collected about them and how it is collected, as well as ask for personal information to be deleted by the business (De Groot, 2019).

75 REPORT Growing concern over consumer data privacy is contributing to debates around best practices for data collection and analysis. To mitigate public concern and plan for an automated future, policymakers can develop policies that encourage responsible data practices from SAV operators as well as develop in-house capabilities for safely storing and managing data. POTENTIAL IMPACTS Independent of the business models, use cases, built environment type of and policies on AVs, the deployment of AVs could have a notable impact on: • Travel behavior, • Environment, • User demographics, • Economics and labor, and • Social equity. The following sections review how AVs could impact travel behavior, the environmental impacts of AVs, the role user demographics could have on AV deployment, the potential economic and labor impacts associated with automation, and the potential impacts—both opportunities and challenges—associated with equitable service. Travel Behavior The impacts of automation on travel behavior are uncertain. Existing roadway capacity may increase due to more efficient operations associated with technology (e.g., closer vehicle spacing known as platooning). Additionally, operators could “right-size fleets,” providing consumers with vehicles sized based on the length of a trip and the number of passengers. However, there is a possibility that AVs could induce demand by making motorized travel more convenient and affordable than personal driving. This could adversely impact congestion (Stocker and Shaheen, 2018). Harper et al. (2016) estimate that AV introduction would increase total annual light-duty VMT by 14 percent, mainly by current adult non-drivers. Likewise, Harb et al. (2018) explored travel behavior shifts due to AV introduction and concluded that AVs allow underserved populations to have more travel freedom. The study predicted an 83 percent VMT increase. The impacts of AVs on congestion will likely depend on whether the vehicles are predominantly shared or privately owned. Public policy, such as pricing and restrictions on ZOVs, could also impact congestion. Soteropoulous et al. (2019) reviewed 37 modeling studies (more than half focused on US regions) and found that most studies indicate that privately owned AVs will increase VMT; reduce public transit and “slow mode” (i.e., active modes of transportation including walking and biking) modal share; and lead to more dispersed urban growth patterns. Conversely, SAVs could reduce congestion by lowering the number of vehicles and parking spaces. Another travel behavior consideration is the impact on mode choice including existing modes of personal vehicles, public transit, and walking. Several studies have attempted to enhance understanding of future mode choice. Using a model based on Austin, Texas, Chen and Kockelman (2016) predicted that a shared fleet of EVs could comprise 27 percent of trips generated and that most of these trips would be at the expense of privately owned vehicles.

76 REPORT Davidson and Spinoulas (2016) studied growth scenarios in 2035 and 2046 and found that active transportation modes would increase market share over time, even as SAV fleets increased. Applying a game-theoretic approach to model interactions between AVs and pedestrians in urban neighborhoods, Millard-Ball (2018) observed that, because AVs will be risk-averse, pedestrians will be able to cross streets more freely. This could lead to a shift toward pedestrian-oriented urban neighborhoods, though possible planning and policy responses are wide-ranging. Sessa et al. (2015) conducted a survey that found P2P SAVs with no pooling could lead to more trips and fewer public transit trips, while a system of SAVs owned by a third-party business or government could complement public transit and draw trips away from private vehicles. The authors assumed that, as automation increases, the ease by which users can switch between modes of transportation will increase, providing first- and last-mile access and reducing the non- monetary costs for using public transit. The communities that travel behavior models are based on may influence predictions about a modal shift. For example, a smaller, auto oriented community (e.g., Austin, Texas) may predict shifts away from the private automobiles, while communities, such as Boston, Massachusetts or San Francisco, California, with more robust public transportation networks, could see shifts away from transit. Emerging studies of MOD services, such as TNCs, tend to support this hypothesis (Erhardt et al., 2019; Gehrke et al., 2018; Henao and Marshall, 2019; Hampshire et al., 2017; Rayle et al., 2016). While the impacts of AVs are uncertain, future mode choice will likely vary based on a variety of local factors including the built environment, service scaling (e.g., fleet sizes), and multimodal connectivity (e.g., the ease in which riders can switch between public transit and an AV). Initial trust in the capabilities of ADS, perceived safety, and perceived privacy risk are also critical factors affecting mode choice. A survey-based study by Zhang et al. (2019) that focused on partially automated vehicles (SAE Level 3) found initial trust was the most critical factor in promoting a positive attitude towards AVs. The automation of electric vehicles could make AVs a more convenient transportation option than privately owned vehicles. This convenience, paired with features, such as self-parking and self-driving, could support the shift away from privately owned vehicles. Travelers may use AVs for point-to-point mobility and/or as first- and last-mile connections to other forms of transit. The use of AVs to fill gaps and offer select services could continue to support the reduction of privately owned vehicles and vehicle purchases. Environment Similar to potential changes to travel behavior, the impact of AVs on the environment is uncertain. Travel behavior and the propulsion systems of the vehicles (e.g., gasoline, electric, fuel cell) can alter the impacts of AVs on the environment. Research has found that the potential emissions reductions benefits of SAVs are usually driven by public policy that promotes EVs and regulates demand (Brown and Dodder, 2019). For example, Greenblatt and Saxena (2015) found that right-sizing the vehicles in an SAV network could reduce per-mile GHG emissions in the range of 63 to 83 percent compared to privately owned hybrid vehicles by 2030. The authors found even more savings with electric AV fleets. However, researchers note that the energy mix used to charge the EVs could reduce any emission savings (e.g., if the energy used to power the car is produced using coal, natural gas, or oil). To preserve SAVs’ capacity for emissions

77 REPORT reductions, renewable energy generation will need to keep pace with increases in demand from electrification (Crute et al., 2018; Patella et al., 2019; Wu et al., 2019). In addition, there is no clear process for sourcing raw materials, recycling, or reusing batteries for EVs (Kendall et al., 2018). This environmental concern may impact the widespread adoption of EVs, which could impact electric AV adoption. Changes in travel behavior could undermine the benefits gained from lower-emission vehicles. Ross and Guhathakurta (2017) found that full automation could result in more energy consumption due to induced travel demand. Likewise, Crute et al. (2018) suggest that the potential convenience of SAVs may generate additional short-range trips that may have been previously completed on public transit, with an active mode, or avoided altogether, resulting in an increase in VMT. Crute et al. (2018) also suggest that, as MOD and AVs decrease shipping costs, total VMT will increase for goods delivery. Additional VMT may also be generated by zero-occupancy vehicles, or empty vehicles traveling between passenger rides (National Academies of Sciences, Engineering, and Medicine, 2019). Millard-Ball (2019) uses a traffic microsimulation model and data from downtown San Francisco, California to predict the impact of AVs seeking out free on-street parking, returning home, or cruising. The results of this model suggest that AVs could more than double vehicle travel within dense urban cores as well as lengthen existing trips (Millard-Ball, 2019). User Demographics Similar to the influence travel behavior and propulsion type have on the environmental impacts of AVs, the demographic makeup of the early adopters of AVs may influence the potential impacts of AVs. Current research suggests that future SAV demographics may be similar to current TNC users. Bansal et al. (2016) surveyed residents of Austin, Texas to explore the acceptance of automated driving technology. They found that “higher-income, technology-savvy males” were the most interested in using AVs. Additionally, the authors found that older licensed drivers were less interested in SAV services. The results mirror the current demographics of shared mobility services, showing higher use by younger, Caucasian, male, and higher-income users (Shaheen et al., 2017a). Menon et al. (2019) surveyed American Automobile Association South members as well as students, faculty, and staff of three University of South Florida campuses about their willingness to give up a household-owned vehicle if SAVs were available. Influential factors included gender, number of household vehicles, current travel characteristics, collision history, and other demographic characteristics (e.g., level of educational attainment). The findings indicated that a wide range of socio-economic factors influence a person’s likelihood of vehicle relinquishment in the presence of SAVs. Age is another important consideration for SAV adoption. Shergold et al. (2016) conducted a literature review on older adults and AVs. The study concluded that older adults would be less likely to embrace AVs than younger users. The authors emphasized that the most important difference was how older adults adopted new technologies. In contrast, Gkartzonikas and Gkritza (2019) reviewed potential user preferences toward AVs through choice experiments conducted between 2012 to 2018. Among their findings, the researchers noted that people over 60 and people between 18 and 25 were the most willing to pay to use AVs.

78 REPORT Studies have documented that mobility declines after peaking between ages 40 and 50. SAVs have the potential to increase mobility and enhance the general quality of life for older adults. More research is needed to understand how older adults will respond to SAVs. It is important to remember, however, that, by the time SAVs are fully scaled, millennial drivers could be part of the older population themselves and may be more likely to adopt SAVs (Shergold et al., 2016). Economy and Labor Automation will also have impacts on the economy and labor market breakdown. Companies are often motivated to invest in automation due to the desire to reduce labor costs. As a result, Level 4 and Level 5 SAV fleets could result in job losses. However, automation could also result in demand for other jobs, such as attendants to assist people with disabilities and older adults and a high-tech workforce to maintain an automated fleet. A simulation study by Groshen et al. (2019) found that AVs could directly eliminate 1.3 to 2.3 million jobs over the next 30 years, depending on various technology adoption and diffusion scenarios. The introduction of AVs could lead to growth in transportation use, new labor inputs for the AV sector, and increased purchases of other goods and services from those who spend less on transportation. A Department of Commerce study estimated that in 2015, 15.5 million (one in nine) US workers were employed in occupations that could be affected, to varying degrees, by the introduction of AVs. This includes 3.8 million “motor vehicle operators” and 11.7 million “other on-the-job drivers.” The Commerce Department estimates that motor vehicle-operator jobs are most concentrated in the transportation and warehousing sector and comprised of a predominately male, older, less-skilled workforce. Other on-the-job drivers use roadway motor vehicles to deliver services or to travel to work sites—such as first responders, construction trades, repair and installation, and personal home care aides—and are mostly concentrated in construction, administrative and waste management, health care, and government sectors of the economy. These workers may be more likely to benefit from greater productivity and better working conditions offered by AVs than motor vehicle-operator occupations (Beede et al., 2017). The USDOT’s most recent assessment of AVs on the transportation system announced that they were working with other cabinet agencies on a comprehensive analysis of the workforce impacts of AVs. This analysis would inform a transition strategy for manual driving-based occupations (USDOT, 2018). In addition to manual driving-based occupations, the USDOT will likely need to research how vehicle automation can impact public transit agency employees. Individuals in roles of drivers, dispatchers, mechanics, etc., may face changing job responsibilities or demand due to vehicle automation. The Eno Center for Transportation (2018) concluded that the benefits of AV deployment outweighed the adverse impacts on the existing workforce. They found that, in a single year, the benefits gained from widespread AV deployment could be greater than the total costs to workers over the next 35 years combined. The study cited potential economic benefits including improved safety, increased commuter productivity, reduced energy consumption, and enhanced accessibility. The report found that the possible costs from unemployment and foregone labor income would start in the early 2030s and increase at a rate of 0.06 to 0.13 percentage points and

79 REPORT peak between 2045 and 2050. The report concluded that the impacts of the 2008 financial crisis were greater than any forecasted job losses from AVs. AVs pose opportunities and threats to existing transportation-related business models. Pütz et al. (2019) found that primary insurers focused on private retail motor insurance face a significant threat to their business model since AVs are likely to substantially lower accident risk exposure. AVs may also have an impact on the revenue-generation models of airports, which rely on fees from parking, ground transportation, and rental cars. A simulation study by Wang and Zhang (2019) found that increased AV market penetration rates caused fewer passengers to choose to park at the airport. The study suggested using alternative fees, such as curbspace access fees, to address lost revenues from the transition to AVs. In a global review of emerging AV regulatory policy, Taeihagh and Lim (2019) found that governments have mostly refrained from instituting mandatory measures in favor of establishing voluntary initiatives and creating working groups to explore AV implications. Social Equity In addition to potential economic and labor impacts, AVs may also enhance access and economic opportunities for underserved communities. However, there can also be impacts on social equity if a population or community bears a disproportionate share of the benefits or adverse impacts of AVs (e.g., lack of services in low-income or minority communities, no SAV fleet parking or recharging in low-income neighborhoods). In a research article evaluating how MPOs are planning for equitable impacts of emerging technologies, Kuzio (2019) categorizes future deployment scenarios for AVs into two main opposing views. In the first scenario, individuals use AVs similarly to taxis or TNCs in a “transportation-as-a-service” model, and car ownership is low. In the second scenario, users retain ownership of vehicles and switch to autonomous modes. This second scenario exacerbates inequity, potentially pricing out disadvantaged populations or forcing them to become public transit dependent. Notably, both scenarios would not lower VMT under current policies. Kuzio (2019) suggests that there is a gap in how MPOs are considering the implications of technology on equity. Likewise, researchers are concerned about the potential for inequity in safety and air quality impacts. For example, the best safety systems are primarily fitted on new luxury vehicles, and it would take time for these features to reach mass-market vehicles (Zmud and Reed, 2018). Dean et al. (2019) reviewed existing literature on public health impacts of AVs and found that, while there is consensus that AVs will improve road safety, the relationship between air quality, physical activity, stress, and other health factors is more complex. The authors indicated that regulatory action and thoughtful planning and design will be significant determinants in the equitable distribution of health impacts. Shaheen et al. (2017) developed the STEPS framework to understand the potential barriers that travelers using MOD may face. This framework can also be applied to understand potential equity issues associated with AVs (both privately owned and shared) (Shaheen et al., 2017a).

80 REPORT Table 16 defines each part of the STEPS framework as well as opportunities and challenges a SAV network may face when addressing social equity concerns. Table 16. SAVs and the STEPS Framework Definition SAV Opportunities SAV Challenges Sp at ia l Spatial factors that compromise daily travel needs (e.g., excessively long distances between destinations, lack of public transit within walking distance) • Public transit agency and first- and last-mile partnerships • Cost-effective SAV service for low-density areas • Land use changes • Higher operating costs in lower-density exurban and rural settings • Limited curbspace for increasing MOD and SAV services T em po ra l Travel time barriers that inhibit a user from completing time-sensitive trips, such as arriving to work (e.g., public transit reliability issues, limited operating hours, traffic congestion) • Dynamic, on-demand transportation service • SAV availability in late nights and early mornings • Wait time and travel time volatility on congested roadways • Unpredictable wait times due to supply fluctuations E co no m ic Direct costs (e.g., fares, tolls, vehicle ownership costs) and indirect costs (e.g., smartphone, internet access, credit card access) that create economic hardship or preclude users from completing basic travel • SAV subsidies for low- income users • Multiple payment options (e.g., cash, card, mobile) • Mobility hubs or SAV vehicles with public wi- fi access • Disruption to existing revenue streams (e.g., parking, traffic violations) • High cost for long- distance/peak-demand trips • Affordability maintenance while providing livable wages Ph ys io lo gi ca l Physical and cognitive limitations that make using standard transportation modes difficult or impossible (e.g., older adults, people with disabilities) • Older adult or child- focused, ADA compliant, SAV services • Voice-activated app features • Maintenance of legacy technology access • Necessity of adequate training if chaperones or chauffeurs are needed for the SAV ride So ci al Social, cultural, safety, and language barriers that inhibit a user’s comfort with using transportation (e.g., neighborhood crime, poorly targeted marketing, lack of multi-language information) • App interfaces that minimize or eliminate sociodemographic profiling • Targeted outreach to low-income and minorities • Network information in the user’s native language • Inclusion of marginalized groups to SAV service • Security provision at unmanned SAV stations, hubs, or pick- up locations

81 REPORT Social equity can be difficult to analyze because several types of equity challenges could impede a user’s access to AVs. For example, current MOD modes often require a smartphone, mobile internet access, and/or a credit or debit card. As a result, these services could exclude access to digitally impoverished, low-income, and unbanked and underbanked users. Additionally, service availability may be limited or unavailable for low-density and rural areas, older adults, and people with disabilities. However, SAVs could create opportunities to enhance access and equity by providing increased mobility options (e.g., lower fares, a greater variety of routes), increased travel speed and reliability, first- and last-mile connections, and expanded coverage to historically underserved users or communities. Legislation and regulation could play an important role in transportation equity by preventing and mitigating technological and access barriers (Shaheen et al., 2017a). Legislation and regulation can also help to ensure that AVs are accessible and do not become a burden on taxpayers and public agencies while benefitting private companies. KEY TAKEAWAYS • AVs are vehicles with some level of automation in their sensory, processing, navigation, and/or communication systems that can relieve a human operator of some operational functions. • AVs are enabled by a variety of technologies working together including systems engineering; sensors and perception; mapping; PNT; communications and security; and human factors. AVs may also have additional technologies called “features” depending on the vehicle manufacturer and intended use (e.g., passenger mobility, goods delivery). • AV deployment will likely be influenced by business models, use cases, and partnerships; the built environment; and policies and regulations. • AV deployment could potentially impact travel behavior, the environment, user demographics, the economy, labor, and social equity. • Current AV pilots are typically testing vehicles or carrying passengers, operating on public or private roads, and using a low-speed shuttle or conventional vehicle.