Shared Automated Vehicle Toolkit: Policies and Planning Considerations for Implementation (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.

6 The convergence of several trends—including shared mobility, the commodification of trans­ portation, digital integration, electrification, and automation—are contributing to the growth of new transportation services, such as MOD, AVs, and ADVs. The following chapters of this toolkit discuss the development of these modes, their predicted deployment, and potential impacts. Mobility on Demand MOD is a concept based on the principle that transportation is a commodity where modes have economic values that are distinguishable in terms of cost, journey time, wait time, the number of connections, convenience, and other attributes. MOD enables consumers to access mobility, goods, and services on­demand by dispatching or using shared mobility, delivery services, and public transportation through an integrated and connected multimodal network (Shaheen, Cohen et al. 2017). Five aspirational attributes can characterize MOD: 1. Commodifying transportation choices into economic terms based on cost, journey time, wait time, the number of connections, convenience, and other attributes. The consumer is making travel decisions based on wait time, travel time, price, etc. These factors are causing commuters to use different modes of travel. 2. Embracing the needs of all users (travelers and couriers), public and private market partici­ pants, and services across all modes—including motor vehicles, pedestrians, bicycles, public transit, for­hire vehicle services, carpooling/vanpooling, goods delivery, and other transpor­ tation services. 3. Improving the efficiency and reliability of the transportation system and increasing the acces­ sibility and mobility of all travelers. 4. Enabling transportation system operators and their partners to monitor, predict, and adapt to changing transportation conditions across the entire mobility ecosystem (network). 5. Maintaining the ability to receive data inputs from multiple sources and provide responsive strategies targeting an array of operational objectives. Mobility as a Service MOD may be supported by other emerging innovations, such as mobility as a service (MaaS). MaaS is an integrated mobility concept emerging where travelers can access their transportation modes over a single digital interface. Although MOD and MaaS have similarities, MaaS emphasizes passenger mobility, allowing travelers to seamlessly plan, book, and pay for a multimodal trip on a Background

Background 7   pay­as­you­go and/or subscription basis. MOD, on the other hand, emphasizes passenger mobility through on­demand access to a variety of transportation modes; this access may be supported by or facilitated through MaaS. Table 2 describes the differences between MOD and MaaS. MaaS can support MOD by providing users with access to a variety of transportation options (e.g., TNCs, micromobility) in the MOD ecosystem. The U.S. DOT defines the MOD eco system as an integrated and multimodal transportation operations management approach that can interact with and/or influence the supply and demand sides of MOD (Shaheen, Cohen et al. 2017). The supply side consists of the professionals, operators, and devices that provide trans­ portation services (e.g., public and private mobility services, fixed­route public transit, goods delivery services, transportation facilities, and information services). The demand side consists of the users of transportation services (e.g., travelers, couriers, consumers, and modal demand). Figure 2 visually represents the MOD ecosystem. The same ecosystem characteristics are likely to be core components in an automated future. Public agencies, developers, and private operators are already deploying AV programs to test use cases for AV technology. Many of these pilots implement low­speed, automated shuttles that can be called on demand or operate on a fixed route. A few commercial services offer on­ demand mobility similar to TNCs. Vehicle Automation In addition to recent transportation developments, such as MOD and MaaS, vehicle automation has the potential to impact transportation networks. AVs are vehicles with some level of automation in their sensory, processing, navigation, and/or communication systems. The level of automation may vary from a single automated function to a vehicle that can operate without a human driver in any environment. AVs that are shared among owners or passengers are referred to as shared auto­ mated vehicles, or SAVs. Privately owned AVs and SAVs are electrified rather than depend on gas­ or diesel­powered engines. Eventually, the majority of AVs are likely to be EVs. In the Characteristic MOD MaaS Description A concept based on the principle that transportation is a commodity where modes have distinguishable economic values. MOD enables customers to access mobility, goods, and services on-demand. An integrated mobility concept in which travelers can access their transportation modes over a single digital interface. MaaS primarily focuses on passenger mobility allowing travelers to seamlessly plan, book, and pay for travel on a pay- as-you-go and/or subscription basis. Payment • Pay per mode. • Pay per trip: Fare payment for multiple modes is integrated. • Pay a subscription fee to a variety of services (e.g., flat fee for bikesharing and carsharing services per month). • Pay-as-you-go. Example In 2017, the Livermore-Amador Valley Transit Authority (LAVTA) in Dublin, California, began partnering with two TNCs and a taxi company to offer residents subsidies on on-demand rides that began and ended in Dublin. The goal of the program is to offer residents more mobility options and reduce costs on underused bus routes (City of Dublin n.d.). The service has since expanded to the surrounding towns (Baum 2020). UbiGo in Northern Europe operates as a transportation brokerage service providing member households a mobility subscription in place of car ownership. The monthly subscription allows households to pre-purchase mobility access in a variety of increments on multiple modes, operating like a multimodal “digital punch card” for a number of transportation services, including public transportation, carsharing, rental cars, and taxis (UbiGo n.d.). Table 2. MOD versus MaaS.

8 Shared Automated Vehicle Toolkit: Policies and Planning Considerations for Implementation SOURCE: Shaheen, Cohen et al. 2017. Figure 2. The MOD ecosystem.

Background 9   future, this can help reduce transportation­related emissions, particularly if AVs are charged using clean energy. However, the availability of charging infrastructure may present challenges to growing and mainstreaming EVs and AVs. The impacts of AVs are uncertain and will likely depend on how they are integrated into the broader transportation network. A variety of automated systems (e.g., sensory, navigation, positioning, communication) func­ tioning simultaneously allow AVs to operate. SAE International, a global mobility standards organization, has established six levels of vehicle automation. These levels of automation define the degree of control needed from the human operator or provided by the vehicle. SAE’s levels of automation are: • Level 0: Vehicles that are not automated, and drivers perform all tasks; • Level 1: Vehicles that automate only one primary control function; • Level 2: Vehicles with automated systems that have full control of specific vehicle functions such as accelerating, braking, and steering, but drivers must still monitor driving and be pre­ pared to immediately resume control at any time; • Level 3: Vehicles that allow drivers to engage in non­driving tasks for a limited time. Vehicles will handle situations requiring an immediate response; however, drivers must still be pre­ pared to intervene within a limited amount of time when prompted to do so; • Level 4: Vehicles that a human operator does not need to control as long as the vehicles are operating in the specific conditions for which they were intended to function; and • Level 5: Vehicles that are capable of driving in all environments without human control. Figure 3 summarizes these levels of automation and design considerations. The safe deployment of AVs may depend upon if they are developed to be fail­safe or fail­fatal. Fail­safe is a design fea­ ture that ensures that if a device fails it will respond in a way that will cause minimal or no harm to the user or surrounding individuals. If a design is not fail­safe, it may fail in a way that results in injury or damage to the user or surrounding individuals. As Level 4 and Level 5 AVs (sometimes referred to as highly automated vehicles or HAVs) become more mainstream, it is possible that a variety of shared mobility services (e.g., car­ sharing, car rental, TNCs, and taxis) may converge to create a new model comprised of SAVs. The introduction of HAVs into the marketplace could offer the potential to increase roadway safety by removing human error in driving. HAVs could also improve infrastructure use by employing technology to reroute vehicles to underserved routes, reduce vehicle spacing, and move goods deliveries to off­peak times. The growth and implementation of HAVs are likely to vary based on differences in the built environments where they are deployed. There may be associated limitations if technologies, such as AVs, are deployed before techno­ logical maturation. These limitations could affect operating areas, operating speeds, passenger size (e.g., not configured for children or smaller riders), and hours of operation. NCHRP Web- Only Document 331: Mobility on Demand and Automated Driving Systems: A Framework for Public-Sector Assessment contains further information on AV technologies and design features. ADV Development Increasing vehicle automation is also contributing to the development of specialized AVs for specific use cases, such as goods delivery, that is, ADVs. ADVs are being designed to serve multiple customers per route and are currently being tested and piloted for last­mile grocery, food, and package delivery. Four types of emerging ADVs include the following. 1. Fully Automated Delivery Vehicle: Fully automated delivery vehicles use Level 4 or 5 auto­ mation and can operate with limited to no supervision by a delivery driver. Level 5 ADVs may be designed to operate without a driver altogether. However, human couriers may be necessary

10 Shared Automated Vehicle Toolkit: Policies and Planning Considerations for Implementation for some delivery use cases including vehicle­to­door delivery (i.e., a courier delivers a package to a building or requires a signature). Fully automated delivery vehicles may be paired with smaller delivery robots to reduce the need for human labor. Privately owned AVs can be used as ADVs to deliver food and packages through peer­to­peer (P2P) delivery business models, similar to current delivery services using privately owned vehicles. 2. Semi-Automated Delivery Vehicle: Semi­automated delivery vehicles, primarily at Level 3 automation, may be able to perform some driving tasks under the supervision of a driver. The delivery driver can also perform some tasks while in transit (e.g., coordinating pickup, updating package delivery). 3. ADV with Lockers: ADVs could also contain lockers that can be used for mobile delivery. Customers could be notified in advance of the exact time of delivery and then pick up their delivery using a unique unlock code from the vehicle locker. 4. Mobile Staging Base: Similar to ADVs with lockers, ADVs could act as a mobile staging base that brings goods to a predetermined point. A human employee then delivers the package from the ADV to the package’s final destination. AV Deployment As vehicles become increasingly automated, they will likely be deployed in different phases. Although analyst predictions vary, AVs will likely require a lengthy transition period before becoming commonplace on U.S. roadways. The transition to AV penetration on the transportation SOURCE: Shuttleworth 2019. Figure 3. SAE International’s levels of automation.

Background 11   network has important implications regarding policy development around privately owned AVs, SAVs, and ADVs. Shaheen, Cohen, and Stocker (forthcoming) developed a four­ phase transition framework that can be applied to passenger vehicles and heavy­duty vehicles (e.g., cargo trucks). Phase 1 outlines present­day vehicle automation technology and pri­ vately owned AV and SAV applications. Phase 2 describes possible conditions as AV technology slowly improves to be able to operate across more envi­ ronments, and automated features become more prevalent in new privately owned vehicle models. Phase 3 represents an important milestone where privately owned AVs and SAVs could see rapid growth and adoption. Phase 3 is the most critical point at which policy deci­ sions will be most urgent and influential. This phase will require safety regulations related to the removal of in­person operators as well as policies aimed at mitigating potential negative impacts (on congestion, emissions, safety, etc.) due to greater uptakes of privately owned AVs and SAVs. AVs at this and subsequent levels will need to comply with the National Highway Traffic Safety Administration’s (NHTSA) Federal Motor Vehicle Safety Standards (FMVSS), or the manufacturer or importer will need to acquire an FMVSS exemption. In addition, if AVs are purchased with Federal Transit Administration (FTA) funding (e.g., to be used by local public transit agencies) they will need to comply with Buy America, a regulation that requires goods purchased with FTA funding to be manufactured in the United States. Phase 4 outlines the proliferation of Level 4+ automation without in­vehicle supervision. Table 3 summarizes possible implications and services that may emerge during each of these four phases considering both privately owned AVs and SAVs. While Table 3 refers mostly to passenger AVs, many predict that freight and goods movement (especially on highways) may be one of the first widespread AV applications (Groshen et al. 2018). SAV Deployment In addition to levels of automation and implementation phases, how SAVs are deployed could alter their use cases and change their impacts. According to Stocker and Shaheen (2018), SAVs are likely to be deployed in one of three models: 1. Business-to-Consumer (B2C): The B2C model most closely resembles carsharing operators. In a B2C model, a vendor owns or leases a fleet of vehicles that are accessible to riders via a subscription membership and/or per­use fee. The vendor typically provides insurance and maintenance for the fleet. The term “business” can include government entities, non­profits, and public­private partnerships. 2. Peer-to-Peer: In the P2P model, a company provides the resources to facilitate the short­term use of a vehicle between a host (vehicle owner) and guest (vehicle lessee). The P2P company typically provides insurance and retains a portion of the rental fee. 3. For-Hire: In this model, customers hire a vehicle on an as­needed basis, either on­demand or through a reservation. Examples of the for­hire model include taxis, TNCs, and microtransit. As SAVs become more common, B2C and P2P business models may converge with for­hire services (e.g., carsharing and TNC services converge as vehicles pickup 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 an SAV network at scale will not require a human driver, there will no longer be a need to distinguish between “for­hire” 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. The NCHRP Web-Only

12 Shared Automated Vehicle Toolkit: Policies and Planning Considerations for Implementation Document 331 contains more detailed information on how business models could impact SAV service characteristics. Figure 4 summarizes some characteristics that may differ in SAV busi­ ness models. Depending on how SAVs are deployed, they can potentially complement, mirror, or compete with services offered today by MOD operators. The antici­ pated use of SAVs is most similar to the MOD modes of carsharing, taxis, taxi sharing, and TNCs (pooled and single trip). Stakeholders can apply and/or adapt strategies for the emergence of AVs to take advantage of the opportunities and address associated challenges. SAVs may be deployed to serve a variety of use cases. Table 4 describes some potential use cases. Phase 1: Present Day (i.e., 2020) Phase 2: Specific-ODD Automation Phase 3: Citywide-ODD Level 4 Phase 4: Proliferation of Level 4+ Pr iv at el y O w ne d A V s Small penetration of Level 2 automation on select new vehicle models (e.g., Tesla Autopilot, Nissan ProPILOT Assist). Level 2 and 3 features continue to roll out slowly in new vehicle models; may see greater penetration of privately owned AVs used for highway driving; potential alteration of infrastructure to support AV use (e.g., development of AV- only zones). Some private ownership of Level 4 vehicles; AV penetration will depend on upfront and operating costs of automated technology; AV retrofit kits could increase private AV ownership rates. Decreasing the cost of AV technology may make ownership possible for a greater portion of the population; private AV ownership may be segmented by land-use context due to cost differences (e.g., AV ownership in suburban/rural areas). Fl ee t o f S ha re d A V s Small but growing number of low-speed SAV pilots and testing (e.g., EasyMile) and SAV pilots and testing with conventional vehicles (e.g., Waymo) in controlled environments. Additional SAV pilots emerge, most likely at low speeds and serving specific use cases and geographical areas (e.g., first- and last-mile to transit services, office parks) in less-controlled environments (e.g., public roads) without onboard drivers. Removal of a human operator in vehicles and introduction of SAV services in major cities and metro areas with high rates of travel demand; potential for vehicle ownership reductions, modal shifts away from public transit, and personal vehicle driving. Shared fleet SAV services gain more ridership and expand to more cities and into some suburban areas; possibility for an even greater reduction in privately owned vehicle rates but also in public transit ridership; alternatively, vehicle ownership and public transit ridership rates could increase. ODD = operational design domains. Table 3. Potential privately owned AV and SAV services during implementation. Vehicle Size Small Vehicles Increasing Vehicle Size Large Vehicles Temporal Attributes On-Demand Flexible-Schedule Fixed-Schedule Spatial Attributes Point-to-Point Flexible Route Fixed-Route Figure 4. Possible SAV service characteristics.

Background 13   Use Case Description Example Pa ra tr an si t S er vi ce The Americans with Disabilities Act (ADA) requires public agencies that provide fixed-route transit to provide accessible transportation services for individuals who live within three-quarters of a mile of fixed-route public transit and are not able to access transit due to barriers, such as cognitive disabilities or physical barriers. However, this service can be costly for agencies to provide and may result in challenges for riders, such as long wait times. SAVs may help reduce costs 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. 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 the current paratransit systems use, SAVs may be able to decrease wait times and trip times for paratransit users. Fi rs t- an d L as t- M ile C on ne ct io ns Travelers may have difficulty getting to or from public transportation (commonly referred to as the first- and last-mile challenge). Public transit agencies are engaging in a variety of partnerships with mobility service providers to bridge these spatial gaps and increase access to public transportation. SAVs may be able to help bridge first- and last-mile gaps by offering connections to transit stops and stations. Fi xe d- R ou te P ub lic T ra ns po rt at io n Many public transit modes operate on a fixed route (e.g., buses, light-rail systems). Fixed routes could present an opportunity to deploy new transportation modes to allow vehicles serving existing routes to be repurposed to serve other areas or increase the frequency of existing routes. SAVs could operate on a fixed route 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. L ow -D en si ty / Sp at ia l G ap -F ill in g Se rv ic es Lower-density built environments may have less frequent transit service and lower transit ridership, which increases the cost of providing public transportation services. Lower ridership and higher operational costs can contribute to lower levels of service for consumers (e.g., longer wait times and fewer routes). To help overcome this challenge, some public agencies are partnering with mobility service providers to offer gap-filling services in lower- density communities. 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 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. O ff- Pe ak /T em po ra l F ill in g C ha lle ng es Providing off-peak or late-night transportation services can be cost- prohibitive for some communities. Additionally, many travelers may not want to wait for infrequent late-night transit service after dark. Public agencies can provide alternative services or options during off-peak hours by partnering with service providers to provide demand- responsive options during periods of lower ridership. 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 both public agencies to provide and riders to own) to fill service gaps. Table 4. Potential SAV use cases. (continued on next page)

14 Shared Automated Vehicle Toolkit: Policies and Planning Considerations for Implementation Appendix A: “Sample Policy Agreements” contains sample policy agreements for offering automated shuttles, gap­filling services, and emergency services. These policy agreements can be found in subsections Appendix A­2: “Sample Policy Agreement for Automated Shuttles,” Appendix A­3: “Sample Policy Agreement for Gap­Filling Services,” and Appendix A­4: “Sample Policy Agreement for Emergency Services.” ADV Deployment As ADVs are deployed they are likely going to be regulated under the same rules as AVs (either by regulation, legislation, or executive order). Existing regulations that guide goods delivery (e.g., when loading zones can be used, compulsory freight weighing) can also be expanded to include ADVs and other Nuro Automated Vehicle Grocery Delivery Service Nuro completed an automated grocery delivery pilot between August 2018 and March 2019 in partnership with Fry’s Food Stores in Scottsdale, Arizona. The pilot demonstrated consumers’ willingness to use ADVs. In April 2019, Nuro and Kroger launched an automated grocery delivery service in Houston, Texas (Hawkins 2019). Although early tests used a safety driver, the backup driver was removed in late 2019. Deliveries used automated Toyota Priuses and cost $5.95. In summer 2020, Nuro began partnering with Walmart to delivery groceries from Walmart stores to residents in the Houston area (Hawkins 2019). In April 2020, Nuro received approval to test their vehicles in California for future delivery services (Crowe 2020). Use Case Description Example C lo se d C am pu s T ra ve l Closed campuses may be expansive and benefit from a transportation system that serves the area. This includes theme parks, resorts, malls, business parks, college campuses, airport terminals, construction sites, downtown centers, real estate developments, gated communities, and industrial centers. SAVs could provide short-distance, point-to- point travel in closed campus environments. E m er ge nc y R es po ns e Responding to an emergency requires coordination across a variety of agencies and systems. In addition, a response can pose safety challenges for responders, particularly if not enough resources are diverted to the scene or the supplies are delayed. In the future, SAVs may provide emergency transportation. One potential use case is an automated ambulance that determines the most efficient path from an emergency incident to a medical facility. U rb an G oo ds D el iv er y Courier network services (CNS)—also referred to as flexible goods, app-based, or on-demand delivery—are apps that 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. This may be particularly useful during times, such as the COVID-19 pandemic, when transit services and access to goods and retailers are reduced. Table 4. (Continued).

Background 15   California Department of Motor Vehicles’ Light-Duty, Autonomous Delivery Vehicle Authorization In January 2020, the California Department of Motor Vehicles (DMV) began reviewing applications for companies to operate ADVs on public streets. The DMV has established requirements for testing with a driver, testing without a driver, and public deployment. Some key requirements include • Certifying vehicles have been tested in controlled environments, • Ensuring that drivers maintain a clean driving record, • Certifying that vehicles can operate under Level 4 or Level 5 automation, • Informing the DMV of the intended operational use, • Ensuring the vehicle meets industry standards, and • Certifying the manufacturer has conducted tests. Source: California Department of Motor Vehicles 2019. Potential Benefits and Challenges of Privately Owned AVs and SAVs If deployed similarly to MOD and integrated well into existing transportation networks, privately owned AVs, SAVs, and ADVs can potentially offer benefits including • Decreased congestion and commute times, • Fewer environmental impacts than traditional transportation modes, • Greater variety of transportation options, • Improved transportation services for vulnerable populations (e.g., older adults, low­income households), • Increased individual mobility, • Lower vehicle ownership costs, and • More access to resources (e.g., medical care, employment opportunities). However, if not properly planned for and integrated, these transportation innovations could result in challenges, such as • Data sharing and privacy concerns, • Exclusion of select demographics (e.g., rural communities, low­income households), • Greater GHG emissions, • Increased vehicle miles traveled (VMT), • Increased use of the energy grid for electric AV charging, last­mile innovations and potentially limit the negative impacts. However, as vehicles become increasingly automated and goods delivery continues to evolve, new regulations may need to be developed and implemented. Some communities, such as Arlington, Texas, and Austin, Texas, have launched ADV pilot programs. In December 2019, California’s Department of Motor Vehicles (DMV) approved the testing and use of light­duty (vehicles under 10,001 pounds) ADVs on public roads (California Department of Motor Vehicles 2019). ADVs will likely require similar considerations as AVs, such as curbspace management and equity considerations (to ensure access to the benefits of ADVs by disadvantaged communities and users with special needs). Appendix B: “Sample Policies” includes an example policy for ADV considerations.

16 Shared Automated Vehicle Toolkit: Policies and Planning Considerations for Implementation • Less efficient infrastructure use, • Loss of jobs and reduction of employment opportunities, • Rider safety concerns, and • Worsened congestion. Potential Impacts of ADVs While ADVs have the potential to increase goods access, they may also result in adverse impacts. Potential undesirable impacts of ADVs, freight vehicles, and other automated goods delivery services include: • Affordability: Fees for last­mile delivery services, such as automated grocery delivery, may not be affordable for select populations (e.g., low­income households) and may require smart­ phone or banking access to use these services. • Accessibility: Last­mile delivery devices may not be accessible for users (e.g., people with dis­ abilities, older adults). Companies developing ADVs should consider using universal design principles to ensure populations can access packages and food from ADVs. • Congestion and Curbspace Management: A growing number of last­mile delivery vehicles may increase curbspace and rights­of­way congestion and require congestion mitigation strategies, such as loading zones, pricing to limit deliveries during peak periods, etc. • Impacts on Pedestrians, Cyclists, and People with Disabilities: ADVs may block bicycle lanes, curbs, ramps, and other ADA access for people with disabilities unless they are explicitly regulated by public agency guidelines and legislation. • Infrastructure Maintenance: Increased use of infrastructure from ADVs, in addition to impacts from existing transportation modes and systems, could require increased mainte­ nance of infrastructure (e.g., roadways, curbs, and loading zones). Considerations for ADV Deployment • Understand the differences in ADV designs when trying to predict and address their potential impacts. • Use ADV pilots as a resource for gaining a better understanding on potential impacts of ADVs. • Employ tools, such as surveys, to better understand the public’s perception and concerns regarding ADVs. • Use strategies, such as policies, to address potential negative impacts of ADVs. • Consider existing ADV regulations when developing or expanding upon ADV regulations. • Ensure that regulations for ADVs are flexible to allow for changes to be made as ADVs develop.