State and Local Impacts of Automated Freight Transportation Systems (2023)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

1   Background Technology advances have made the dream of self-driving vehicles a near reality, and automated vehicles (AVs) of all kinds have captured the public imagination. Self-driving cars seem to be accepted as the wave of the future, despite intense scrutiny of every crash and incident. Ordinary family cars are increasingly equipped with automated safety features, and some are now capable of extended “hands-off” operation on designated routes. High- profile companies such as Uber and Google have invested in self-driving taxi technology. Small sidewalk robots (SADRs) delivering meals are a regular sight in some cities and near some college campuses. Remote-controlled aerial drones are in widespread use for photog- raphy and recreation, and automated drones seem like a logical extension. The U.S. DOT has issued a series of guidance and policy publications, most recently, Ensuring American Leader- ship in Automated Vehicle Technologies: Automated Vehicles 4.0 (1). Many public and private sector observers believe autonomous freight is coming as well. Transportation planners are being urged to anticipate regular use of freight automated vehicles (FAVs) in years, not decades. SADRs and drones are delivering meals in real-world tests, and press articles predict regular use of autonomous trucks in the near future. As a field, AVs are technologically advanced but remain in the testing and demonstration phase. FAV demonstrations to date include SAE Level 3-4 semi-tractor/trailer combinations on limited-access highways; closely monitored SADRs and other urban delivery vehicles; and aerial delivery drones with and without automation. The research team found that the FAV state of the art is at SAE Level 3, with most vehicles at SAE Level 0 or 1. Automated safety features are widely available in passenger vehicles, and they are becoming more common in freight vehicles; however, they are present in only a small portion of the existing fleet. FAV deployment will depend on successful introduction into the supply chain and the complex milieu of state and local highways and streets. Integration into the supply chain will determine where FAVs can increase efficiency and provide private benefits. Integration on public roads and sidewalks will determine where they can operate safely and deliver public benefits. State and local impacts of autonomous trucks and urban delivery vehicles are thus largely unknown and unexplored, and state and local agencies need information, insight, and guid- ance. The extensive research and public debate over the future of passenger automated vehicles (PAVs) are not paralleled by work on FAVs. State and city authorities are being alternatively buffeted by demands to move faster and grasp opportunities, and cries to slow down and ensure safety after every adverse incident. To manage a smooth transition to a more automated freight future, public policy must balance the expected benefits of FAV operation with the additional burden placed on state planners and local jurisdictions. Research to date has been focused primarily on vehicle technology, leaving gaps in knowl- edge of infrastructure, legal, and planning needs. S U M M A R Y State and Local Impacts of Automated Freight Transportation Systems

2 State and Local Impacts of Automated Freight Transportation Systems Scope This study focused on state and local implications of FAV deployment so FAVs can operate beyond the demonstration phase, and so state and local agencies can make the infrastructure, policy, and procedural changes required to reap the potential benefits. The research objectives were to (1) develop a decision framework for use by state and local entities, and (2) document those requirements in a final report. The research, framework, and report cover the expected public sector impacts of automation in the following: • Over-the-road freight automated vehicles (OTR FAVs), namely, self-driving trucks. • Urban delivery vehicles using sidewalks and streets, namely sidewalk automated delivery robots (SADRs). • Unmanned aerial vehicles (UAVs) or delivery drones. Rail, air, and waterborne freight transport are also increasingly automated, as are warehouse and distribution center operations. Those operations are not on public right-of-way and do not have the same state and local implications. These have been addressed in the appendices. FAV Deployment Little is known about the eventual deployment of FAVs, the timing of that deployment, or their commercial viability. In general, FAVs will be commercially deployed in significant numbers where four main factors overlap: operational design domain (ODD), net private ben- efits, public acceptance, and public infrastructure and regulation (see shown in Figure 1). Figure 2 summarizes deployment scenarios for OTR FAVs, SADRs, and UAVs. Deploy- ment will likely expand stepwise rather than gradually because the four major factors in Figure 1 expand coverage and the overlap increases in response to future trigger events or developments. In each case, deployment is likely to progress from a limited ODD similar to current testing and demonstrations to broader use as ODDs, commercial viability, and public acceptance expand. • OTR FAVs are likely to be initially deployed in an intermodal handoff scenario, with Level 3-5 operation on long highway trips between points where operation is transferred Figure 1. FAV deployment factors.

Summary 3 to human drivers for first-/last-mile access. The next deployment level is likely to be Level 4, door-to-door use on specific routes defined by ODDs and mapping, with eventual expan- sion to near-universal access if technically possible. • SADRs are also likely to be deployed, if commercially feasible, on limited routes and in limited urban areas, much as they are being tested at present. Opening wider access to SADRs will depend on their ODD, public acceptance, and the commercial feasibility of longer trips or serving less dense markets. • UAV deliveries are now being tested without automation in a few niche markets, which is where they are likely to be first deployed if commercially viable. They may also be used to extend human delivery routes to serve outlying customers. As with SADRs, wider deploy- ment will depend on an extended ODD, public acceptance, and commercial success. The scenarios have some features in common: • Each technology might be restricted to designated routes or areas by public regulation. The regulation could take the form of route-by-route approval, prohibitions on some routes, or a combination. • At the other extreme, each technology might eventually enjoy unlimited access, with deployment choices made by private sector operators and customers. • Each technology will require a navigation solution to enable both macro-navigation (way- finding) and micro-navigation (steering, lane choice, parking, etc.), and those solutions may require a mix of infrastructure improvements, mapping, and technology advances. • Each technology will require, at a minimum, remote monitoring by exception, placing the technology on the boundary of automation Levels 3 and 4. Table 1 summarizes the research team’s findings on potential supply chain roles for FAVs. These potential roles include those within the capabilities of existing FAVs and those that would require new but foreseeable capabilities (such as handling returned goods via SADRs or UAVs). Truck drivers and delivery persons perform multiple functions besides driving. Deployment of unoccupied FAVs will depend on finding new ways to perform those functions, including delivery at destination; customer interaction; fueling or charging; vehicle inspections; load monitoring; hazardous materials (HAZMAT) handling; and coping with incidents. Curb access will be a major issue in the potential deployment of self-driving trucks or other Figure 2. FAV deployment scenarios by mode.

4 State and Local Impacts of Automated Freight Transportation Systems street-going FAVs in urban delivery service. UAV access presents a different set of problems where urban rooftop access may be impractical or unsafe. In current parlance, there must also be a “business case” for deploying FAVs, and that busi- ness case has not yet emerged. Expectations of commercial FAV viability rest on assumptions regarding net labor savings, safety benefits, capacity improvements, or increased customer satisfaction that have yet to be demonstrated or quantified. For OTR FAVs, the potential private benefits may include labor savings, safety improvements, fuel savings, alleviation of the driver shortage, increased capacity, and improved service. Against these potential ben- efits, operators will have to weigh the additional costs of new vehicles; remote monitoring systems and staff; remote retrieval and servicing; and additional terminal staffing. Commer- cial viability of SADR and UAV operations will depend on their economics relative to other pickup and delivery options; customer acceptance and willingness to pay; development of a cost-effective remote monitoring system; and the nature and size of the accessible market. Conceptual Timelines FAV development and deployment timelines are necessarily speculative, given the cur- rent state of development. Table 2 displays a conceptual timeline for OTR FAV deploy- ment, public investment, and public benefits. This timeline could shift markedly as the noted trigger events occur sooner or later. For OTR FAVs, the longevity of freight vehicles, particularly diesel trucks, implies a lengthy fleet turnover period relative to passenger vehicles. The availability and usability of remote monitoring for Level 3-5 operations are critical for significant labor savings. Truck fleet operators may continue to incorporate Level 1-2 features and automated safety system (ASS) capabilities into their fleet, as the timeline suggests, without ever proceeding to Level 3-5 operation. Perhaps the most signifi- cant feature of the timeline is the long time required for OTR FAVs to become a significant presence in the overall freight fleet and thus yield the anticipated safety, road congestion, and environmental benefits. Table 3 presents a conceptual timeline for SADR deployment and impacts, which assumes the following: a business case emerges for SADRs in regular delivery use; cities and other Generic Movement Type Generic Shipment Size Parcel or Small Shipment LTL Truckload (1) Raw Materials from Source OTR FAVs (2) Processed/Manufactured Goods from Production to Distributors or Customers OTR FAVs OTR FAVs OTR FAVs (3) Finished Goods from Producers or Distributors to Resellers or Retailers OTR FAVs OTR FAVs OTR FAVs (4) Deliveries to Multiple Customers SADRs OTR FAVs OTR FAVs OTR FAVs (5) Delivery to Individual Customers SADRs UAVs OTR FAVs (6) Used Goods, Returns, Recycling, Waste SADRs UAVs OTR FAVs OTR FAVs OTR FAVs LTL: less than truckload. SADRs: sidewalk autonomous delivery robot. OTR FAV: over-the-road freight automated vehicle. Table 1. FAV supply chain roles.

Year 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 Conceptual OTR FAV Trigger Events Level 5 Intro. AV Fleet Share at 25% AV Fleet Share 50+% AV Fleet Share 70% AV Share of Vehicle Fleet 0% 0% 0% 0% 0% 0% 0% 1% 1% 1% 1% 1% 2% 2% 3% 4% 6% 10% 14% 20% 26% 33% 41% 48% 55% 62% 70% Base/Autopilot Level 0-2 Deployment Intermodal Handoff Level 3-5 Deployment Route-Specific Level 3-5 Deployment Universal Level 3-5 Deployment ASS/ADAS Usage and Benefits Level 3/4 Operation on Limited Access Freeways Level 3/4 Operation on Surface Streets Level 3/4 Operation on Rural Roads Level 5 Operation Infrastructure Investment Public Safety Benefits Public Congestion Benefits Remote Level 4 Monitoring and Control Allowed Common V2I Communi- cation Initial Level 3-5 deployment in intermodal operations over limited access freeways Increased AV Sales and Adoption Reliable All- Weather Operations Work Zone and Incident Solution Work zone solution and V2I communications facilitate deployment beyond freeways Final stage of Level 3-5 deployment with all technology in place Gradual penetration of Level 2, ASS/ADAS technology Public congestion benefits with share over 25% Level 1-2 ASS/ADAS penetrates fleets with or without Level 3-5 AV functions Start of Commercial AV Sales Level 5 operations once Level 3/4 ODD becomes "everywhere" Public safety benefits grow with ASS/ADAS usage and Level 3/4 operations Safety benefits grow with fleet penetration Level 3/4 ODD expands with infrastructure investment in all-weather operations Progressively greater infrastructure and communication investment to keep up with ODD expansion Declining investment Level 3/4 ODD expands with infrastructure investment and V2I communications Common V2V Communications Introduced Rapid Adoption and Fleet Penetration Level 3/4 Intermodal or Route-Specific Ops. ADAS: automated driver assistance system. ASS: automated safety system. AV: automated vehicle. V2I: Vehicle-to-infrastructure. V2V: Vehicle-to-vehicle. Table 2. Conceptual OTR FAV deployment timeline.

6 State and Local Impacts of Automated Freight Transportation Systems jurisdictions permit SADRs in at least some instances; and many-to-one monitoring systems emerge in 2024 to 2026. Unlike the OTR FAV case, there is no existing fleet for SADRs to penetrate, but there is also no basis for predicting the extent of future SADR use. Table 4 presents a conceptual timeline for delivery UAV deployments. This assumes an emergence of a viable business case for regular UAV deliveries and FAA permission for autonomous UAV operation beyond visual line of sight (BVLOS), neither of which is cer- tain. As with SADRs, there is no reliable basis for predicting the extent of future delivery UAV deployment. State and Local Impacts The research yielded insights into potential public benefits, requirements, and opportu- nities that are explained in terms of responses to basic public sector questions regarding the future of FAVs: What net public benefits can be expected? The expected public benefits of FAV deploy- ment include increased safety, improved transportation efficiency, reduced congestion, and some positive environmental impacts. These may be offset by negative environmental impacts and employment losses. Expected safety improvements are the most commonly cited reason for public support of AV deployment. The implicit or explicit assumption is that removing the human element through automation will significantly reduce accidents and fatalities. Besides reducing vehicle- to-vehicle (V2V) accidents, automated monitoring of bridge and tunnel clearances; size and weight restrictions; and HAZMAT routing should reduce vehicle-infrastructure accidents and incidents. There is a critical distinction to be made, however, between the potential ben- efits of automated safety system/automated driver assistance system (ASS/ADAS) and the incremental benefits of the automated driving system (ADS) in Level 3-5 driverless operation once ASS/ADAS benefits have already been achieved. The research team was unable to locate any estimates of either the benefits of well-developed and widely used ASS/ADAS features or the incremental safety benefits of driverless operation. While Level 2 vehicles will become more common and Level 3-4 vehicles will likely become available, the benefits of automa- tion will depend on how often and where those capabilities can be used. Year 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 Conceptual SADR FAV Trigger Events Limited Route Deployment Open Access Deployment Infrastructure Investment Public Congestion Benefits Street congestion vs. sidewalk congestion Navigation Solution SADRs tightly restricted by place, time, and number SADR use limited only by safety exceptions Possible sidewalk, crosswalk investment Many-to-1 Monitoring SADR: sidewalk autonomous delivery robot. Table 3. Conceptual SADR deployment timeline.

Year 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 Conceptual UAV Delivery Drone Trigger Events Emergency/Public Safety Deployment Niche Deployment Delivery Route Extension Deployment Universal Deployment Infrastructure Investment Public Street Congestion Benefits Progressively greater infrastructure and communication investment for UAV ATC Street congestion benefits vs. airspace congestion Emergency/Public Safety use beginning with manual control, gradual transition to autonomy Universal use requires autonomy, many-to-1 monitoring, and ATC Development of UAV ATC Niche deployments of autonomous UAVs following demonstration uses Extension deployment requires many-to-1 monitoring for scale economics FAA Permission for BVLOS and Autonomous UAV Operations Delivery Route FAA Permission for Many-to-1 Monitoring ATC: air traffic control. Table 4. Conceptual delivery UAV deployment timeline.

8 State and Local Impacts of Automated Freight Transportation Systems Expected improvements in freight transportation efficiency from OTR FAVs depend on assumptions regarding labor savings, 24/7 operations, and more economical OTR opera- tion due to automated control. These expectations implicitly assume widespread Level 4-5 operations, which have not yet been achieved and which may not be achievable. SADRs and UAVs might offer efficiency advantages over individual deliveries via auto, bicycle, motorbike, or foot, yet they are probably less efficient than route-based delivery service or multi-stop errand running by individual customers. Congestion reductions will depend on multiple factors, including the degree to which ASSs and AVs of all kinds penetrate the passenger and freight fleet operating on public roads, and the effectiveness of those systems in improving traffic flow and road capacity. Congestion reductions may be offset by increased vehicle activity encouraged by automation. Neither of those factors has been researched in any depth to date, and the extent and timing of congestion reductions remain speculative. The extent of environmental benefits from FAV deployment will depend on the relative size of positive (reduced environmental impact) and negative (increased environmental impact) outcomes. Positive impacts may come from improved OTR FAV efficiency through automated controls; substitution of SADR or UAV operations for inefficient individual vehicle trips; and reduced street and highway congestion (i.e., reduced idling, improved fuel economy). Negative impacts may stem from increased vehicle miles traveled (VMT) due to intermodal handoffs instead of direct trips, and additional delivery trips induced by convenient SADR or UAV systems. The extent and balance of the positive and negative impacts are unknown at present. FAV deployment is likely to reduce employment, but the nature and extent of net job losses are still highly uncertain. Labor savings are the primary motivation for Level 4-5 OTR FAV deployment, so successful deployment of self-driving OTR FAVs will likely result in a net loss of jobs. The reduction in driver jobs will likely be offset by additional jobs at terminals and elsewhere, but the net impact will likely be negative. Both SADRs and UAVs will require many-to-one remote monitoring to achieve commercial viability and would therefore reduce employment compared with the same deliveries being made by humans. The impact will be greater to the extent that SADRs and UAVs are substituted for human delivery trips, and less if SADR and UAV services generate new trips. Under what conditions should the public allow FAV operation? From all the research done to date, safety emerges as the overwhelming criterion for permitting FAV operation on public streets, on public sidewalks, and in public airspace. Fundamentally, states and localities should, and likely will, allow FAV operations that are safe and publicly accept- able. These are likely to develop in parallel with PAV operations. Public agencies do not determine whether operation is beneficial to the individual or organization operating the vehicle, or creates net public benefits. Vehicles and drivers that have met safety, licensing, and regulatory requirements are allowed to use public streets and highways. This study pre- sumes that once appropriate requirements are set and met, FAVs will likewise be allowed progressively broader street and highway access. The concept of safety, as displayed in the conceptual decision framework, goes beyond minimizing the potential for collisions, injury, and damage to include, for all FAV modes, the following: • Vehicle equipment, standards, inspection, licensing, and cargo securement. • Driver/operator training, testing, licensing, insurance, safety records, hours of service (HOS), and reporting. • Law enforcement, work zone, and incident management. Most of these issues are common to FAVs and PAVs. The most significant difference is that FAVs, if they reach SAE Levels 4-5, will ordinarily be unoccupied by either driver or

Summary 9 passenger. Many of the decisions faced by the public sector and considered in this research are driven by the lack of an in-vehicle operator to navigate, recognize problems, inspect the vehicle and its cargo, and interact with the environment as needed. FAV mode-specific requirements include the following: • OTR FAV lane control, routing, access, restrictions, and other operational rules related to the ODD. • SADR “rules of the sidewalk.” • UAV air traffic control (ATC). As shown in the conceptual decision frameworks, these and other requirements must be adapted for application to FAVs. The frameworks thus constitute a public sector “checklist” for allowing FAV deployment. For unoccupied OTR Level 4-5 FAV operations and for all freight UAV and SADR operations, some form of remote operations monitoring will also be required, with appropriate public safety provisions. Public acceptance is required for FAV deployment in public, almost by definition. There are no technical criteria for public accep- tance nor can it be readily quantified. Yet, as a practical matter, state, regional, and local governmental bodies are subject to public pressure over contentious issues and are unlikely to permit widespread FAV deployment in the face of public hostility. What must state and local agencies do to prepare for FAV deployment? Physical infra- structure requirements for FAV (and PAV) deployment appear to be relatively modest due to: (1) industry development that stresses adaptation to existing infrastructure, and (2) ODD limitations to route and circumstances where infrastructure is suitable. The lack of consistent, effective pavement striping and markings is currently a barrier to widespread deployment of PAVs and FAVs using camera or sensor navigation. Besides limiting deploy- ment of Level 2-4 operations, striping and marking shortfalls limit the effectiveness of ASSs, and thus the potential public and private safety benefits of vehicle automation even if autonomy is never achieved. Lack of state-to-state and local standardization of signage, markings, geometrics, and traffic control devices is also a barrier to widespread FAV (and PAV) deploy- ment. The need to standardize, however, is being reduced as AV developers find ways to cope with a wider range of circumstances. Considerable attention has been paid to the potential advantage of communications with and between AVs. The current industry approach implies that public agencies will not have to provide V2V, V2I, or vehicle-to-everything (V2X) com- munications infrastructure, but also implies that vehicle movement and road condition data expected to flow from such communications will not be forthcoming. Policymakers are also concerned about potential cybersecurity breaches and the increased data and control vul- nerability of AVs. Design and operation standards for FAVs (as with PAVs) may require a high grade of complexity to ensure the protection of personal information; surveillance regulation; and compliance with physical, network, and information requirements. Liability in AV accidents remains an open question. The emergence of FAVs operating on roadways without a traditional driver may require structural change in the liability model to account for the different roles of remote monitors or operators versus system and vehicle designers. These issues are likely to be tested in cases involving self-driving autos. FAV law enforcement issues include adherence to traffic laws; crash reporting; criminal activity; first responder and law enforcement training; operational responsibility; permit conditions enforcement; inspections; vehicle identification and vehicle response to atypical road conditions; emer- gency vehicles; manual traffic controls; work zones; load limits; clearances; and HAZMAT restrictions. These concerns stem primarily from the lack of onboard human operators. How can state and local agencies realize the potential public benefits of FAV develop- ment and deployment? FAV deployment will be bound by ODD. Public efforts to expand the potential ODD through improved pavement markings, standardized traffic controls,

10 State and Local Impacts of Automated Freight Transportation Systems and other initiatives would facilitate FAV (and PAV) deployment. Moreover, many of these same initiatives would yield safety benefits from ASS/ADAS systems short of Level 3-5 oper- ations, and from manually driven vehicles. As with other transportation regulation and traffic laws, states and localities tend to adopt successful approaches and language from other jurisdictions. Development of model laws, statutes, regulations, and ordinances could speed the process for both FAVs and PAVs, and the Uniform Vehicle Code (UVC) will need to be adapted accordingly. Some reports suggest public education efforts to increase acceptance of PAVs and FAVs. These suggestions presuppose the potential for significant net public benefits from safety improvements, efficiency, and congestion relief. The existence and size of those benefits have yet to be established, and as noted elsewhere, there are reasons to dis- count the most optimistic perspectives. If focused on safety, public education efforts might be comparable with campaigns for seat belt use or against texting while driving. Decision Framework The relevant high-level state and local considerations in FAV deployment are diagrammed in Figure 3. For each proposed FAV deployment, the public sector can choose whether to allow it on existing infrastructure, improve the infrastructure first, limit the scope of deploy- ment (the ODD), or deny the deployment. These key factors are shown as a checklist on the right of Figure 3. If these conditions have been met, the deployment should logically proceed. If not, the deployment should be limited, the infrastructure improved, or the deployment denied. For each mode (OTR FAV, SADR, or UAV), the research team developed a two- or three- step decision framework beginning with deployment prerequisites common to PAVs and FAVs, moving to FAV-specific issues common to all deployment scenarios, and scenario- specific issues. The OTR FAV framework is shown as an example in Figure 4. BCA: benefit-cost analysis. Figure 3. High-level FAV deployment evaluation.

Note: subsequent figures zoom into each stage. Public Figure 4. OTR FAV decision framework.

12 State and Local Impacts of Automated Freight Transportation Systems In general, the first step addresses factors such as safety, regulations, and public accep- tance that must be in place before any PAV or FAV deployment can take place. The second step addresses issues such as remote monitoring, cargo securement, and trip inspections that must be resolved before FAV deployment but are not barriers to PAV deployment. The final step addresses factors that must be addressed to permit progressively wider FAV deployment, such as the ability to operate in all weather or negotiate urban curb access. This stepwise process can guide state and local agencies in their planning for FAV deployment and in reacting to specific deployment proposals.