URBAN DENSITY AND AUTONOMOUS VEHICLE: 

COMPARATIVE INSIGHTS FROM CHINA AND THE 

UNITED STATES 

 A Dissertation  

Presented to the Faculty of the Graduate School 

of Cornell University 

in Partial Fulfillment of the Requirements for the Degree of 

Master of Science 

by 

Shi Yan 

December 2025



 

         

 

©2025 Shi Yan



 

         

 

 

ABSTRACT 

 

This paper examines the deployment of autonomous vehicles (AVs) across different urban 

densities, focusing on the cases of China and the United States. While much existing research 

focuses on technological breakthroughs or national policies, I am more interested in 

understanding how urban form influences the direction, pace, and outcomes of AV deployment. 

 

To explore this question, I develop a framework linking policy instruments, technological 

strategies, and urban density. I apply this framework to a series of city-level case studies, 

including Beijing, Suzhou, Chengdu, New York, Pittsburgh, Chicago, San Francisco, and 

Phoenix, as well as rural areas such as Shanxi and suburban Arizona. I also draw on the social 

construction of technology theory (SCOT) to understand the role of different social groups in 

the development and implementation of AV technology. 

 

The analysis reveals that population density, in addition to being a contextual factor, is a key 

variable. In densely populated cities like Beijing and New York, the complex road 

environments require strong coordination between the deployment of AVs and public 

infrastructure and regulations. In China, the dense population forces the government to 

prioritize dense infrastructure, such as vehicle-to-everything (V2X). In contrast, American 

cities like San Francisco and New York often rely on private industry to drive innovation, 

resulting in autonomous vehicles with more robust onboard intelligence. However, how to 

effectively integrate them into urban systems remains an unresolved issue. 



 

         

 

 

The differences are equally striking in less densely populated cities like Suzhou and Pittsburgh. 

The Suzhou government actively plans to integrate autonomous vehicles into its industrial 

parks, even developing China's first smart highway. In Pittsburgh, early industry-university-

research collaborations helped initiate autonomous vehicle testing, but a lack of sustained 

public policy support has slowed its wider adoption. This suggests that moderately dense cities 

may have the potential to combine flexibility with infrastructure investment, provided 

governance capacity is aligned. 

 

In low-income and rural regions with relatively low population density, both China and the 

United States are gradually leaning toward a more car-centered strategy. Because the costs of 

building advanced infrastructure are high while the expected returns remain limited. For 

example, in Arizona, companies like Waymo have begun operating autonomous vehicle 

services in suburban communities. Meanwhile, in rural China, supported by simplified policies 

and relatively practical technology, companies are actively experimenting with autonomous 

vehicles in areas such as logistics and last-mile delivery. 

 

Overall, this paper demonstrates that while population density plays an important role, political 

institutions and governance models often override or reshape these spatial constraints. For 

example, China’s top-down governance structure enables many medium-sized cities to take a 

proactive role in planning and investing in AV-related infrastructure. In contrast, American 

cities tend to rely more on market-driven experiments led by private companies. As a result, 



 

         

 

even cities with similar population densities can have very different trajectories for autonomous 

vehicle development in both countries.



 

 iii 

 

BIOGRAPHICAL SKETCH 

Shi Yan is a master’s student in the Regional Science program in the Department of City 

and Regional Planning at Cornell University. His research examines the interaction between 

emerging technologies and urban development, with a particular focus on how autonomous 

vehicle deployment varies across different urban densities and governance systems in China 

and the United States. 

Before coming to Cornell, he completed his undergraduate studies at Pennsylvania State 

University and majored in Geography, where he developed an interest in geographic 

information system and spatial analysis. During his graduate study, he has also gained 

experience with GIS technologies, data analysis, and comparative policy research. 

 

 

  



 

 iv 

 

ACKNOWLEDGEMENTS 

I am deeply grateful to my advisor, Professor Stephan Schmidt, for his generous 

guidance, patience, and encouragement throughout the dissertation writing process. His 

advice prompted me to focus more on the contrasting political systems of China and the 

United States, which provided me with unexpected insights. I am also deeply grateful to my 

dissertation committee member, Professor John Carruthers, for his thoughtful feedback and 

constructive suggestions that helped me refine my argument. 

 

I am also grateful to the Department of City and Regional Planning at Cornell University, 

and particularly the Regional Science Program, for providing an intellectually stimulating 

environment and the opportunity to learn from colleagues from diverse academic and cultural 

backgrounds. I am especially grateful to my classmates and friends in the program, whose 

discussions, encouragement, and friendship made my studies both fruitful and memorable. 

 

Finally, I would like to express my deepest gratitude to my family. Their unwavering 

support, love, and encouragement gave me the strength to pursue my graduate degree abroad 

and complete this dissertation. This journey would not have been possible without their 

unwavering belief. 

  



 

 v 

 

TABLE OF CONTENTS 

Abstract 

Biographical Sketch .......................................................... iii 

Acknowledgements .......................................................... iv 

Chapter 1: Introduction .................................................. 1 

Chapter 2: Literature Review ......................................... 2 

 2.1 Policy-Technology-Density Framework 

o 2.1.1 Policy Dimension 

o 2.1.2 Technological Dimension 

o 2.1.3 Density Dimension 

Chapter 3: Analytical Framework and Its Application .......... 8 

Chapter 4: Comparative Case Study (United States vs. China) .... 14 

 4.1 Densely Populated Areas 

o 4.1.1 Policy Environment 

o 4.1.2 Technological Adaptability 

o 4.1.3 Urban Density and Development Impacts 

 4.2 Intermediate Density Areas 

o 4.2.1 Policy Environment 

o 4.2.2 Technological Adaptability 

o 4.2.3 Urban Density and Development Impacts 

 4.3 Low-Density and Rural Settings 

o 4.3.1 Policy and Technological Adaption 



 

 vi 

 

o 4.3.2 Impacts of Low Urban Density on AV deployment 

Chapter 5: Summary Table ...................................... 36 

Chapter 6: Discussion .................................................. 36  

Chapter 7: Conclusion ................................................ 38 

Chapter 8: Policy Recommendations ................................ 40 

References .................................................................... 43 

 



 

 1 

 

 

1. Introduction 

 

As early as 1925, only 40 years after Karl Benz invented the first truly modern car, an 

autonomous car designed by Francis Houdina came into existence. By receiving radio signals 

from the vehicle following, the vehicle accomplished the starting, steering, braking, and 

acceleration without the driver control (JALOPNIK, 2017). However, with the limitations of 

computer power and sensors, despite the efforts of scientists and engineers, no breakthroughs 

in autonomous driving technology have been achieved in the 55 years since then. At 80s, with 

the “NavLab” car invented by Carnegie Mellon University and "VaMoRs" launched by 

Mercede-Benz cooperated with Munich Federal Armed Forces University, they prove to the 

world the systematic application of computer vision and multi-sensor fusion technology in 

autonomous driving (Engineer.com, 2016). Entering the 21st century, autonomous driving 

technology has ushered in new development opportunities. Breakthroughs in computer 

computing power, the advent of LiDAR, and the application of AI end-to-end models have 

brought autonomous driving technology one step closer to realizing the vision of large-scale 

application and reshaping transportation methods. Based on the data collected by IDTechEx, 

as of 2024, leading companies such as Waymo, Cruise, and Baidu Apollo have deployed more 

than 2,000 L4 automation level taxis, with a cumulative test mileage of tens of millions of 

miles. Baidu operates 500 driverless taxis in Wuhan and plans to expand to 1,000 by the end 

of 2024. Moreover, one of the leading companies, Tesla, collected real road data through more 

than 1 million mass-produced vehicles, with a cumulative mileage of billions of miles by the 

end of 2024 (Tesla Inc, 2025).  



 

 2 

 

This paper explores the qualitative differences in the formulation of autonomous driving 

policies and the direction of technological development at different urban densities. With the 

insight of cities in the United States and China, I compare their policy framework, technology 

adaptability, and development impacts at different urban densities. Three types of cities are 

evaluated systematically: a) Densely Populated Areas (Beijing & Shanghai vs. New York City 

& San Francisco); b) Intermediate Density Areas (Suzhou & Chengdu vs. Pittsburgh & 

Chicago); c) Low-density or Rural Areas (Yangquan vs. Phoenix suburbs). 

 

The analytical framework of this study adopts the "policy-technology-density" approach. 

Under the perspective of the social construction of technology (SCOT), the complex 

interactions between policy choices, technological capabilities, and density-related urban 

conditions are analyzed. Specifically, the framework considers three interacting dimensions: 

(1) Policy Toolbox: regulatory approaches, subsidies, and pilot initiatives; (2) Technology 

Adaptability: assessing the capabilities, infrastructure complexity, and cost feasibility of 

autonomous vehicles (AVs) in different urban environments; and (3) Density Response 

Thresholds : studying how population, land use patterns, and transportation infrastructure 

capacity affect the feasibility and effectiveness of autonomous vehicle deployment. 

 

2. Literature Review: AV Integration 

 

Witnessing the booming development of the autonomous driving industry, what impact will it 

have on urban development? Supporters argue that the widespread use of self-driving cars can 



 

 3 

 

improve road safety by reducing accidents caused by human error and reduce emissions 

through smoother driving and electric automation (nyc.gov, 2024). According to academic 

estimates, if self-driving cars are widely used in cities, they can reduce urban greenhouse gas 

emissions by about one-third (Makahleh, 2024). These potential advantages make the 

widespread use of self-driving cars a major concern in densely populated urban areas with 

heavy traffic, frequent accidents and pollution. Therefore, how to integrate autonomous 

vehicles (AVs) into urban transportation networks has been one of the most popular research 

directions in the field of urban transportation in recent decades. AVs can perceive the 

environment and navigate without human intervention, thus improving safety, reducing 

emissions and improving efficiency in urban spaces (Fagnant & Kockelman, 2015).  

Due to the fact that AV technology is an emerging industry, which presents some regulatory 

challenges to society, cities and countries are developing policy frameworks to guide this 

transition. AV integration requires appropriate policies to manage regulations for AV testing, 

traffic safety requirements, infrastructure investment to support AV technology, and 

coordination with broader and long-term urban planning goals. As the major countries, both 

geographically and economically, China and the United States have made different attempts at 

AV’s integration. 

 

In the United States, the approach to AV policy is decentralized and often city-specific. As for 

now, there is no national law mandating how AV should be integrated while state and local 

government take the lead in regulation with the guidance provided by federal agencies. 

According to the State of AV 2024 published by Autonomous Vehicle Industry Association, as 



 

 4 

 

of early 2025, 35 U.S. states have enacted laws regulating AV and several others have issued 

executive orders or are actively considering legislation.  

 

In China, national and local governments are actively promoting the deployment of AV as part 

of large-scale smart city initiatives. This is a top-down process: guided by the national-level 

strategy formulated by the central government, local pilot breakthroughs are made to drive the 

coordination of various supply chains and diversification of scenarios. For example, Beijing 

and Shanghai have implemented large-scale AV pilot projects, including vehicle-road 

cooperative (V2I) communication systems and dedicated AV test areas, leveraging the 

government's large investment in smart road infrastructure (Li et al., 2021).  

 

There is still a gap in systematically studying how different urban density environments affect 

AV integration, even though existing research has extensively explored autonomous vehicle 

(AV) technologies, deployment strategies, and their pervasive impacts on cities. This research 

gap highlights the need for a more comprehensive analytical framework that captures the subtle 

interactions between policies, technological capabilities, and urban density. The subsequent 

chapters will fill this gap by introducing a "policy-technology-density" analytical framework 

to compare cities in the United States and China. 

 

2.1 Literature Review: Policy-Technology-Density Framework 

 

The Social Construction of Technology (SCOT) framework proposed by Pinch and Bijker in 



 

 5 

 

the last century argues that instead of driven solely by technological logic, technological 

development is also shaped by social groups, values, and interactions (Lynch & Hu, 2022). In 

the context of autonomous vehicles (AVs), advances in AV technology in terms of design, 

regulation, and usage reflect the influence of stakeholders such as policymakers, industry 

engineers, consumers, and other social groups. Recent research on AVs emphasizes that the 

development of the AV industry should not only focus on the technical field, but also extend to 

the field of social values, and calls for examining AV development through the lens of SCOT 

rather than viewing its application as a purely technical or deterministic process (Waltermann 

& Henkel, 2024). Innovation in the AV industry is fundamentally inseparable from policy and 

regulatory orientation, public expectations, and cultural attitudes. To systematically analyze 

these interacting forces, this paper adopts a three-part coordinated framework, the 'Policy-

Technology-Density' synergistic framework.  

 

2.1.1 Policy Dimension 

 

Policy plays a key role in shaping the trajectory of AV integration. Everything from safety 

standards and testing protocols to subsidies and infrastructure investments are either 

constrained or driven by policy. California has become a pioneer in the AV industry, not only 

because of the dominant of its computer software and electronic components industries, but 

also because of its early policy support for AV. Extensive pilot projects by multiple companies 

in San Francisco's dense urban landscape have made it one of the most active AV testing sites 

in the United States (Stocker & Shaheen, 2017). Comprehensive AV trials have been conducted 



 

 6 

 

early in California, providing a solution to the technical and social problems of high-density 

urban environments (Brodsky, 2016). In contrast, New York City has taken a more cautious 

regulatory stance, not allowing strictly regulated AV testing until 2020. Unlike San Francisco’s 

flat urban plan, New York City is known for its uniquely dense and complex urban environment, 

with high pedestrian traffic and a dense street grid. Beijing, in particular, has established an 

extensive test area covering hundreds of kilometers of urban roads in Yizhuang, equipped with 

sensors and advanced communication technologies (Liu & Wang, 2021). Shanghai has also 

promoted the popularization of AV by issuing licenses to companies such as Baidu and AutoX 

to operate autonomous ride-sharing services in clearly defined areas that provide a policy-

driven integration approach (Li et al., 2021). 

 

 

Figure 1 High-level Autonomous Driving Demonstration Zone in Beijing 

Source: (UNECE) 

 

2.1.2 Technological Dimension 



 

 7 

 

 

Technology adaptability is another dimension that highlights the interaction between AV 

capabilities, infrastructure complexity, and economic feasibility. The success of AVs depends 

on their ability to cope with the complexity of urban environments, from high-density urban 

cores to sprawling suburbs (Fraedrich et al., 2019). AV technology development in the US 

focuses on vehicle self-sufficiency, using onboard sensors and artificial intelligence to navigate 

existing roads, rather than smart infrastructure. The “single-car” autonomous driving approach 

fits the US urban development model, where AVs are expected to be driven more on highways 

and suburban streets, so new infrastructure construction is minimal (Sever & Contissa, 2024). 

China’s AV technology roadmap prioritizes connectivity, reducing the reliance on vehicle-

based technology by counting on smart city infrastructure. Urban pilots connect AVs with 

intelligent transportation systems and dedicated 5G communication networks. Researchers 

have focused on big data-connected AV which can relieve the burden of individual vehicles 

navigating chaotic urban environments and are networked with smart roads and traffic lights 

(Yue et al., 2024). This makes that China’s AV designs often include vehicle-to-everything 

(V2X) capabilities by default, and deployment priorities in complex urban areas tend to favor 

autonomous taxis and autonomous buses, as centrally managed fleets of AVs can ease 

congestion (Ding et al., 2022).  

 

2.1.3 Density Dimension 

 

When people begin to consider how and where the AV should be deployed and integrated, 



 

 8 

 

urban density is one of the factors that first comes to mind. Urban density significantly affects 

traffic patterns, infrastructure requirements and navigation complexity, making it an important 

variable for planning AV applications (Fraedrich et al., 2019). It is challenging even for human 

drivers to drive in high-density urban areas because the dynamic environment is so 

sophisticated - streets are packed with cars, buses, bicycles, and pedestrians that handling dense 

traffic and frequent interactions at intersections requires advanced capabilities and 

infrastructure support for AV. For example, a mature traffic light system, dedicated AV lanes, 

or ultra detailed digital maps are needed to ensure AV can process safely in a busy city center 

(Makahleh, 2024). In contrast, it is simpler for an AV to operate in low-density or highway 

environment since it focuses more on high-speed safety features rather than driving on streets 

with dense pedestrian traffic (Othman, 2022). Policymakers must tailor AV integration 

strategies to urban environments, as what works on quiet suburban roads may not work on 

crowded downtown avenues. Legislative frameworks need to be diverged to population 

density—large cities may introduce special rules that are not applicable to small towns, such 

as designating "zero-emission zones" or integrating AV services with public transportation 

(Kok, 2023). Thus, urban density should be a key consideration and a central theme when 

planning for AV deployment.  

 

3. Analytical Framework and Its Application 

 

This section will elaborate on the “Policy-Technology-Density” analytical structure and 

explain how it will guide the subsequent comparative case studies between the United States 



 

 9 

 

and China. The framework is established with three interrelated dimensions which are policy 

toolbox, technology adaptability, and density response threshold.  

 

Chart 1: Analytical Framework 

 

3.1 Policy Toolbox 

 

Government policies offer a range of tools to influence the development and integration of AVs. 

Scholars have outlined a range of policy influences and interventions for autonomous mobility. 

They emphasize the need for proactive governance to maximize benefits and mitigate risks 

(McAslan, 2021). In practice, many jurisdictions have introduced specific rules for AV testing 

and operation, such as setting safety standards and licensing requirements, while some special 

jurisdictions maintain a relaxed environment to attract industry to experiment. In addition, 

fiscal incentives are another tool for governments. AV research, infrastructure investment, and 

early applications can be incentivized by strategic subsidies or tax credits. Well-designed 

AV Technology

Policy Toolbox 

Rules that support or 
standardize the AV;

Subsidies for infrastructure or 
technology development.

Technological Adaptability

Prerequisites for AV 
Technology; Cost-

effectiveness of deployment 
strategies

Urban Density
Population density; land-use 

pattern; transportation 
network capacity



 

 10 

 

subsidies can accelerate the transition to the "dark ages" of early low market penetration by 

incentivizing continued innovation and us (Luo, 2018). Another key policy tool is to promote 

pilot projects and demonstration projects. Cities around the world have launched pilot projects 

for AVs (such as the Beijing Yizhuang test site mentioned earlier) to understand the feasibility 

of the technology in real-world applications. Pilot projects will allow private AV developers to 

test vehicles on public roads. This will then provide local authorities with insights into the 

impact of AVs on safety, traffic or fairness and allow them to formulate regulations or 

infrastructure plans accordingly. However, these experiments must be carefully coordinated 

with public goals, as they risk marginalizing broader planning discussions about the future of 

mobility if they are simply viewed as technical exercises (Cugurullo, 2021). 

 

3.2 Technological Adaptability 

 

The second dimension focuses on how technical complexity and cost interact with urban 

characteristics. Autonomous driving systems must cope with the quality of infrastructure, street 

geometry, and unexpected traffic behavior in different urban areas. Early deployment is often 

limited to well-mapped areas with favorable conditions, because cities with orderly traffic and 

clearly marked roads can reduce sensor uncertainty, which is conducive to the initial operation 

of AVs. Therefore, the speed of the promotion of AVs depends to some extent on the ability of 

technology to handle the complexity of specific urban environments. Cost and infrastructure 

are also key factors in the adaptability of technology. Significant investments in vehicle 

technology, high-definition maps, vehicle-to-road communications, or dedicated lanes will lay 



 

 11 

 

the foundation for large-scale deployment of AVs, such as designing a consistent lane marking 

system or new communication protocols (Bahrami, 2023). On the one hand, AV developers try 

to improve algorithms to adapt to local driving regulations and restrictions; on the other hand, 

governments and planners may need to install sensors or 5G coverage to support reliable AV 

operation. The high cost of autonomous driving technology is more easily accepted in terms of 

alleviating public transportation labor shortages or improving the safety of dangerous routes. 

Therefore, technical feasibility is closely related to urban conditions. Without supporting 

infrastructure and controllable complexity, even strong policy interventions may have little 

effect. 

 

3.3 Density Response Threshold 

Density response threshold measures how urban density affects the impact and feasibility of 

AV deployment. The urban density is composed by several aspects of urban form which are 

population, land use, and transportation network capacity. 

 

3.3.1 Population Density 

 

Population density is usually measured as the number of residents or jobs per unit area. High-

density cities concentrate many potential passengers and destinations within a certain area, thus 

facilitating shared transportation modes and frequent public transportation services (Heinrichs, 

2016). AV deployment is likely to focus on enhancing collective mobility. Because dense 

passenger flow ensures sufficient demand in a limited area, deploying fleets of autonomous 



 

 12 

 

shuttles or robotaxis to supplement public transportation becomes a preferred option. In 

contrast, in low-density suburban or rural areas, residents rely more on private cars due to 

scattered development (Heinrichs, 2016). To fill the gaps in public transportation coverage, 

autonomous driving technology may mainly appear in the form of personal driverless cars or 

on-demand services rather than in the form of high-capacity shared fleets. Similar to traditional 

transportation, there may be threshold effects in the deployment of AVs. Below a certain density, 

it becomes challenging to operate a cost-effective shared autonomous driving service due to 

longer distances and fewer trip connections (European Environment Agency, 2025). Thus, this 

paper examines these thresholds, recognizing that population density can promote or limit the 

type of AV deployment that is practical in a particular area. 

 

3.3.2 Land-Use 

 

In high-density cities, the concentration of housing, jobs, and amenities can reduce the need 

for long-distance car trips and make people more likely to choose alternative travel options 

such as walking, biking, and public transportation (Heinrichs, 2016). AVs can perfectly 

function as last-mile connections at transportation hubs or as autonomous shuttles in the city 

center. In contrast, in medium- and low-density cities, longer trips and a more car-dependent 

lifestyle driven by dispersed or segregated land use patterns can increase people's consideration 

of the advantages of AVs for long-distance travel. In fact, depending on how they are managed, 

AVs may further promote urban sprawl or promote urban density. Autonomous driving can 

make driving easier and make people more likely to live farther away from the city center, 



 

 13 

 

thereby expanding the coverage of metropolitan areas. One study predicts that if AVs are 

popular in the United States, the urban sprawl area of the Dallas-Fort Worth Metropolitan Area 

will increase significantly by 68% (Moore et al., 2020). On the other hand, the spread of 

autonomous driving technology can also support higher-density development by reducing the 

need for parking spaces and improving land use efficiency in core areas. Existing parking 

spaces can be transformed and connected to AV facilities, making urban spaces more compact 

and efficient. Therefore, land use configuration and policy directions will determine whether 

the deployment of AVs will exacerbate urban sprawl or be consistent with a more sustainable, 

denser urban form. 

 

3.3.3 Transportation Network Capacity 

 

The third density-related factor essentially refers to the ability of the road infrastructure to 

handle the volume of traffic. The introduction of AVs into urban traffic flows will have complex, 

nonlinear effects on traffic grid performance. Traffic engineering studies have shown that at 

low AV penetration, a small number of AVs mixed with a large number of human-driven cars 

will reduce effective capacity and increase congestion, because AVs tend to maintain larger 

safety gaps and drive more cautiously in front of unpredictable human drivers (Li, 2025). 

However, this situation will improve in the future as AV market penetration increases. Once 

the number of AVs reaches a critical level, estimated to be around 50% to 60% of traffic flow, 

efficiency benefits will accumulate through coordinated driving, reduced headway, and 

platooning (Liu, 2024). At that point, urban road capacity may be significantly improved, even 



 

 14 

 

in high-density roads to relieve congestion. This threshold phenomenon means that the traffic 

benefits of AVs are highly dependent on density. Large, congested cities will see improvements 

only after AVs reach a certain level of penetration, while networks with less traffic in small 

cities can achieve smoother traffic even with a small number of AVs. In addition, the process 

of repositioning empty vehicles that have just completed service may lead to increased vehicle 

travel. Unrestricted AV use may lead to additional vehicle miles travel and weaken public 

transportation ridership (McAslan et al., 2021). To ensure that AVs enhance rather than block 

urban mobility at different densities, both infrastructure capacity upgrades and demand 

management policies, such as smart traffic lights or dedicated AV lanes (as mentioned before), 

will be needed. 

 

The following chapter will apply this three-dimensional framework to case studies of Chinese 

and American cities of varying densities, taking a close insight of how policy, technology, and 

spatial factors interact. 

 

4. Comparative Case Study (United States vs. China) 

 

4.1 Densely Populated Areas 

 

For the densely populated areas, I select Beijing, Shanghai from China and San Francisco, New 

York from United State as the case study. Based on the recent demographic data, all four cities 

possess large contiguous contiguous 1ௗkm² grid‑cells each contain ≥ௗ1ௗ500 inhabitants perௗkm², 



 

 15 

 

and the contiguous cluster’s total population is ≥ௗ50ௗ000. 

  

City (core district) Average 
Population 
Density (per 
km²) 

Total 
Population 
(2025) 

Source 

Beijing 5,299 22.6 million World Population Review 

Shanghai 2,059 30.5 million World Population Review 

New York City 11314 8.8 million U.S. Census Bureau 

San Francisco 7,299 0.83 million U.S. Census Bureau 

 

Table 5.1: Population Density of Selected Cities in Densely Populated Areas 

4.1.1 Policy Environment 

 

Chinese Mega-cities benefit an aggressive top-down policy pushing the AV industry. China's 

central government has issued nationwide guidelines and pilot projects to regulate the orderly 

deployment of AV. In Beijing and Shanghai, local governments have strongly supported the 

testing and services of AV by setting up dedicated areas, issuing licenses, and investing in 

infrastructure. By early 2024, Beijing had opened more than 1,160 kilometers of public roads 

for AV trials, and 18 companies had approved 384 AV for trials (The People’s Government of 

Beijing Municipality, 2024). Shanghai has similarly designated four testing areas and 

approved the deployment of 794 AVs on more than 2,000 kilometers of roads for pilots 

(Xinhua News, 2024). Both cities now allow limited commercial autonomous taxi services. 

For example, Beijing decided in mid-2023 to allow fully autonomous taxis (previously, each 



 

 16 

 

autonomous taxi would be equipped with a human safety driver for safety reasons) to operate 

and charge in specific areas (Luo, 2023). This policy has benefited from a series of successful 

supervision pilots since 2019, as regulators have sufficient confidence in the success of the 

pilots. Shanghai plans to introduce Robotaxi in the busy Pudong district and has currently 

authorized several companies to provide such services, such as Baidu, AutoX, and Pony.ai. 

Overall, China's densely populated cities practice multi-level coordinated governance. First, 

the national ministries will formulate strategic goals (to achieve large-scale production of 

AVs by 2025), and then the municipal government will introduce local regulations that suit 

local conditions and promote the establishment of AV innovation pilot zones. 

 

There are no federal regulations for the deployment of AVs in the United States, so states and 

cities need to make their own. More than 60% of the U.S. autonomous taxi deployments are 

in California, which has become the center of the U.S. AV industry. San Francisco's policy is 

relatively open. The California Department of Motor Vehicles (DMV) and the Public Utilities 

Commission have developed a licensing process for AV testing and commercial operations. In 

2024, Waymo and Cruise received licenses to operate 24-hour paid autonomous taxi services. 

San Francisco also became the first city in the United States to allow fully driverless fleets to 

operate commercially. Mayor Lurie has also publicly stated that AVs will be an important part 

of San Francisco's downtown revitalization plan (San Francisco Office of Mayor, 2025). He 

claimed that AV access will improve visitor flow to shops, theatersz and restaurants and 

complement Muni, taxis and bikes. In New York City, state law has set up strict barriers. In 

the early years, AV testing required police escorts. The mayor and the Department of 



 

 17 

 

Transportation introduced a licensing system that requires all autonomous test vehicles to be 

equipped with safety drivers, who must receive professional training and be ready to take 

over at any time (New York City Office of Mayor, 2024). In addition, companies applying for 

testing need to submit historical test records including collision data and disengagement rates 

and prove that their technology can cope with the high density of pedestrians and traffic in 

Manhattan. These strong measures reflect New York City’s dense and complex street 

environment and the policy emphasis on safety and minimizing disruption. 

 

4.1.2 Technological Adaptability 

 

The main technical approaches of China and the United States in densely populated urban areas 

differ significantly. China's deployment emphasizes the vehicle-road cooperation model, using 

vehicle-to-everything (V2X) and smart infrastructure to enhance vehicle autonomy. These 

mechanisms reduce a lot of pressure on AV operating in complex traffic environments. In 

Beijing's Haidian District and Yizhuang pilot area, roadside sensors and 5G communication 

networks transmit real-time data to AVs, improving their perception of traffic and pedestrians 

(Luo, 2023). The Ministry of Transport has explicitly promoted the construction of "smart 

roads". AI cameras, IoT devices and high-precision maps are embedded on highways and city 

streets to support the operation of AV. This reduces the burden on on-board systems by 

providing redundancy and broader situational insights. However, due to the traffic complexity 

of high-density cities, test vehicles in Beijing and Shanghai are generally equipped with 8-12 

lidars, combining millimeter-wave radars with cameras to improve the recognition accuracy of 



 

 18 

 

complex traffic participants. For example, the Baidu Apollo RT6 model is equipped with 8 

Lidars, covering 360° without blind spots (Gasgoo, 2022). Even in dense traffic and complex 

road geometry, China's self-driving taxis can rely on on-board sensors and "smart city" 

infrastructure. 

 

In contrast, most of the AV projects in densely populated cities in the United States follow the 

core of single-vehicle intelligence, relying on on-board sensors and AI algorithms to achieve 

complex environmental perception. In San Francisco, Waymo and Cruise vehicles rely mainly 

on lidar, radar, cameras and on-board artificial intelligence, and rarely rely on real-time input 

from infrastructure. For example, Waymo vehicles are equipped with 13 cameras, 4 lidars and 

6 radars, and multi-sensor fusion is used to improve the recognition accuracy of dense targets 

such as pedestrians and non-motor vehicles. Tesla insists on a pure visual solution, optimizing 

the real-time decision-making ability of urban streets through the FSD (Full Self-Driving) 

system. Some American cities have piloted networked traffic lights or dedicated lanes, but these 

are only limited demonstrations. Chicago, for example, plans to install 10 connected 

intersections as a pilot to aid future AV applications (CDOT, 2021).  

 

The technology stack of AVs in American cities is basically self-sufficient and capable of 

driving on unpredictable city streets. One reason is the lack of ubiquitous V2X deployment in 

the United States; another reason is that regulators prefer technology-neutral infrastructure. As 

a result, China's densely populated cities tend to favor the vehicle-road cooperation model, 

while American cities rely on highly automated vehicles that can handle densely populated 



 

 19 

 

environments with only onboard systems. The Chinese approach may be able to scale faster in 

complex areas by reducing the cost per vehicle and improving safety through network 

monitoring. The American approach prioritizes innovation in vehicle artificial intelligence, 

which has been proven in California's rigorous testing system, but its large-scale deployment 

may be slow if the supporting infrastructure is lacking. 

 

4.1.3 Urban Density and Development Impacts 

 

In these densely populated cities, the introduction of AVs has already begun to influence travel 

behavior and urban planning, although the potential has not yet been fully developed and 

utilized. AV are being regarded as an integral part of smart city projects by policymakers in 

China's mega-cities. They believe that AV can be useful in improving traffic congestion and 

reducing private car ownership in city centers. Moreover, AV should be used as a supplement 

to public transportation rather than a substitute at this stage (Li, 2022). For example, in 

Guangzhou, 50 autonomous buses are deployed on 10 fixed routes connecting major 

transportation hubs such as railway stations and airports; these routes have served more than 1 

million passengers before mid-2023 (Ministry of Transport of the People’s Republic of China, 

2024). Travelers are willing to incorporate autonomous shuttles into their daily transportation. 

In Wuhan, Baidu's Apollo Go autonomous taxis have 500 vehicles, about 1% of the city's total 

taxi amount (Feng, 2024). They are so popular that human taxi drivers have even launched a 

petition to limit their number. In densely populated Chinese cities, people are undergoing a 

gentle shift in travel patterns and are not averse to AVs. Beijing’s draft 2023 AV regulations 



 

 20 

 

even seek feedback on using self-driving cars for urban mobility options such as car rentals 

and ride-hailing, suggesting planners are actively considering how to integrate shared AV fleets 

into the urban fabric. 

In the United States, densely populated urban centers have been more cautious about 

integrating AVs, so the visible impact on cities has been limited so far. In San Francisco, 

Waymo and Cruise AVs, as of 2025, this service still accounts for only a very small part of the 

city's overall travel. Due to the extremely complex traffic environment caused by urban density, 

the problem of AVs stopping unexpectedly often occurs. For example, parking near the Vice 

President Harris warning area and blocking traffic during a music festival (Police1, 2024). The 

cruise was also revoked due to frequent accidents, including dragging pedestrians. In 2024, it 

laid off 50% of its employees and turned to L3 assisted driving research and development, 

resulting in its Robotaxi service almost withdrawing from the market, further weakening the 

coverage of autonomous driving in San Francisco (Hawkins, 2024). The reliability of 

autonomous driving is doubtful, so people are calling on city officials to adjust how and where 

AVs operate. In addition to these early operational issues, land use or urban form in San 

Francisco and New York City have not yet changed significantly due to AVs. Despite this, urban 

planners have not given up studying future scenarios for the popularization of AVs (Silva, 2022). 

Depending on the deployment direction, whether the popularization of AVs may exacerbate 

urban sprawl or may also adapt to higher urban density. If AVs are primarily used as private 

transportation, reducing the cost of commuting, this could lead to increased urban sprawl, as 

people may be willing to live farther from work centers (Booth, 2024). In other case, assuming 

AVs are deployed as robotaxis, the number of private cars and parking needs in cities could 



 

 21 

 

decrease, which could promote denser development in urban cores and more pedestrian-

friendly streets over time (Li, 2025). Currently, US cities are in the early stages of pilot projects 

and have not yet made major infrastructure construction or zoning adjustments for AVs. New 

York City explicitly positions its emerging AV program as integrating AVs into existing urban 

safety and transportation frameworks rather than as a catalyst for reshaping the city. 

 

4.2 Intermediate Density Areas (Suzhou, Chengdu vs. Pittsburgh, Chicago) 

 

For the intermediate density areas, I select Suzhou, Chengdu from China and Pittsburgh, 

Chicago from United State as the case study. Intermediate Density Areas are composed of 

continuous 1 square kilometer grid units, each with a population density of ≥300 people/km², 

and the total population of the agglomeration is between 5,000 and 50,000 people. However, 

the entire urban area does not fully meet the density and coherence of the "City Center". 

 

City (core district) Average 
Population 
Density (per 
km²) 

Total 
Population 
(2020) 

Source 

Suzhou 2,100 12.3 million City Population 

Chengdu Core: 6,200 
Rural: 1,000 

20.9 million City Population 

Pittsburgh Core：2,100 
Rural: < 1,200 

0.23 million U.S. Census Bureau 

Chicago Core: 4,600 
Rural: 1,200 

0.94 million U.S. Census Bureau 

 

Table 5.2: Population Density of Selected Cities in Intermediate 



 

 22 

 

Chengdu and Suzhou have large urban and suburban structures, but the population and building 

density is a multi-center distribution with high core urban areas and medium and low peripheral 

areas. Since large industrial areas, open new areas, and satellite towns have not formed a 

complete ultra-high-density continuous belt, they should be classified as intermediate urban 

areas. 

 

The density of the core district of Pittsburgh and Chicago meets the standard of the “City 

Center”, but they should be classified as “Intermediate Density Areas”as the suburbs are less 

dense and not closely integrated with core district. 

Medium-density cities offer a different context for the deployment of AVs. In these settings, 

where urban forms consist of dense city centers and sprawling suburbs, policy approaches often 

need to balance innovation with local needs. 

 

4.2.1 Policy Environment 

 

China's intermediate density cities are actively participating in the national autonomous driving 

program and often establish themselves as innovation centers to attract investment. Secondly, 

cities implement policy directives by building advanced infrastructure. After the national 

government launched the "intelligent vehicle infrastructure integration" pilot project, Suzhou 

actively responded and took advantage. In 2023, Suzhou opened China's first smart highway 

designed specifically for AV applications and equipped with holographic perception technology. 

Real-time health monitoring and traffic anomaly detection are achieved through distributed 



 

 23 

 

fiber optic sensing, Lidar and other equipment, providing high-precision perception support for 

AVs (Chen, 2023). The Suzhou Municipal Government has also set up AV pilot areas in 

different areas such as some tourist areas, such as Jinji Lake, and high-speed rail new towns, 

in order to test autonomous shuttles and logistics delivery vehicles. The pilot covers more than 

100 application scenarios such as smart travel, freight, and sanitation. Municipal authorities 

attach great importance to the pilot project and therefore receive support from local subsidies 

and partnerships (People.cn, 2022). In 2024, its intelligent vehicle networking industry will 

reach a scale of 45.14 billion yuan, gathering nearly 600 companies (Suzhou Municipal 

People’s Government, 2024). Chengdu, the capital of Sichuan Province, also launched a self-

driving taxi pilot in its high-tech industrial park in 2022. The Chengdu Municipal Government 

initially deployed 8 Apollo Go cars in an area of 10 square kilometers and allowed Baidu to 

provide online taxi services to the public during daytime. Chengdu has incorporated 

autonomous driving into the smart community planning of the suburban "Future Science City" 

to promote the integration of vehicles, roads and clouds. For example, in cooperation with 

telecommute operators to improve 5G vehicle networking facilities, 16 L4 Robotaxi put into 

operation in Longquanyi District provide free services on two routes with a total length of 30 

kilometer (People’s Daily Online, 2023). This is consistent with Chengdu's broader urban 

vision, which clearly regards AVs as part of the blueprint for high-tech sustainable communities. 

In all these cases, the policy means of China's medium-sized cities include the establishment 

of special zones, public-private partnership pilot projects, and the inclusion of AVs in urban 

development plans (People’s Daily Online, 2023). By 2024, China has even expanded the pilot 

project to 20 cities (many of which are second-tier cities, including Suzhou, Chengdu, Wuhan, 



 

 24 

 

etc.) to promote the construction of standardized autonomous driving infrastructure and cloud 

platforms (Zhang, 2024). The pilot requires the integration of vehicles, roads, cloud computing 

and other technologies, and the exploration of commercial operation models to build 

standardized AV infrastructure and AV operation models. This ensures that cities such as 

Suzhou and Chengdu develop in sync with national standards while conducting local pilots. 

 

In the United States, cities with medium population density have taken a more decentralized 

and exploratory policy approach. Pittsburgh is a notable example. Pittsburgh's leadership hopes 

to promote local economic development by introducing high-tech and actively supported AV 

testing in the mid-2010s without federal involvement. In 2016, as one of the first cities in the 

United States to cooperate with the city government and companies in autonomous driving, 

Pittsburgh cooperated with Uber's Advanced Technology Group to allow Uber's AVs to be 

tested on Pittsburgh streets accompanied by safety drivers (BBC, 2016). Pittsburgh was 

therefore designated as an AV testing site by the U.S. Department of Transportation and has 

attracted a number of AV startups and research centers. Many mid-sized cities in the United 

States often lack coherent AV pilot policy goals and are not connected to large-scale 

transportation planning (McAslan, 2021). Pittsburgh's early AV testing was not closely 

integrated with public transportation or transportation goals, and there were some accidents, so 

some people criticized the city for conducting experiments privately without clear public 

interests (SCIAM, 2017). It has since enacted AV testing regulations and participated in state 

government initiatives. Pennsylvania has supplemented its then-incomplete policies with new 

guidelines and voluntary safety commitments. Chicago's approach is more cautious. Chicago’s 



 

 25 

 

2020 Transportation Strategic Plan calls for exploring AV pilots and installing connected 

vehicle infrastructure at some intersections, but concrete deployment efforts have been minimal 

to date (CDOT, 2021). Chicago did approve a limited delivery robot pilot and is considering 

autonomous shuttles for specialized uses like running on dedicated lanes to and from O’Hare 

Airport, but these are still in the proposal stage (CMAP, 2019).  

 

In general, mid-sized cities in the U.S. are taking a more relaxed approach to policy tools—

forming working groups, issuing guidelines, and collaborating on small pilots—rather than 

taking a prescriptive, mandatory approach like China. This difference reflects resource levels 

and risk tolerance. A city like Pittsburgh can relatively quickly approve AV trials by private 

companies but scaling it up to a comprehensive citywide service requires state support and 

coordination with public transit agencies, which is still in its infancy. 

 

4.2.2 Technological Adaptability 

 

In both countries, intermediate density cities are proving grounds for perfecting autonomous 

driving technology, but the balance between vehicle capabilities and infrastructure support 

differs. The way China’s second-tier cities adopt autonomous driving technology is often 

promoted as part of smart city upgrades. Cities such as Suzhou and Chengdu follow the national 

model and place a high priority on infrastructure-enabled autonomous driving. For example, 

Suzhou’s smart highways create a controlled environment where connected AVs (CAVs) can 

receive high-fidelity data from the road, such as real-time traffic conditions and impending 



 

 26 

 

hazards (Gasgoo, 2023). A favored technology path is for infrastructure upgrades to precede or 

accompany the rollout of new technologies for AVs themselves, ensuring that AVs can operate 

with the help of intelligent traffic management systems even outside of densely populated 

metropolitan areas. In addition, given the industrial characteristics of these cities, China’s mid-

sized cities often focus on logistics and freight AV. For example, in eastern China, Zhoushan, 

a port city slightly smaller than Suzhou, has launched a pilot for autonomous trucks at its port 

terminal, while other cities such as Chongqing’s Yongchuan District are testing L4 autonomous 

trucks on 20 square kilometers of roads (CMRA, 2024) (Xinhuanews, 2024). As an industrial 

center itself, Suzhou is likely to benefit from autonomous freight transportation on its newly 

built smart highways. From a technical perspective, this suggests that vehicle-to-infrastructure 

(V2I) communications and cloud coordination are integral to China’s AV systems in a medium-

sized environment – AVs are designed to be part of a network, whether carrying passengers or 

cargo. The Chengdu pilot project, while starting with basic self-driving taxis, coincides with 

Baidu’s efforts to build a 5G-enabled “intelligent driving project” in the city that aims to 

eventually support both self-driving buses and self-driving taxis through a central cloud control 

platform (Spencer, 2021). 



 

 27 

 

 

Figure 1 Smart Highway in Suzhou (source: cnbeta) 

 

In the United States, autonomous technology deployment in medium-density cities has been 

more independent and with relatively limited driver requirements. Pittsburgh's trials primarily 

involved test vehicles equipped with advanced onboard kit that can map the city's streets. 

Pittsburgh's hilly terrain and seasonal weather conditions provide an ideal testing environment 

for AVs to validate technology and help improve vehicle software, but there is little use of the 

infrastructure that is available in Chinese cities (Pennsylvania Autonomous Vehicle Policy Task 

Force, 2016). During Uber's years of testing, no special sensors were installed on Pittsburgh's 

roads. Instead, the city used existing traffic control systems, with the city government providing 

traffic signal data to testers (Patel,2016). Chicago and its surrounding areas have a strong 

freight industry, and we have also seen some autonomous technology for truck transportation. 

Some companies have tested semi-autonomous trucks on highways in Illinois, and nearby Ohio 

and Michigan have corridors for testing truck fleets. These projects use road connections, such 



 

 28 

 

as some vehicle-to-vehicle (V2V) communication between trucks but are not as sophisticated 

as China's city-wide sensor networks (Wisniewski, 2019). In terms of passenger transportation, 

the application of American technology in medium-sized cities usually comes in the form of 

low-speed shuttles or online ride-hailing pilots on selected routes. For example, May Mobility 

already operates autonomous shuttle services in mid-sized urban centers such as Grand Rapids, 

Michigan, and Arlington, Texas, using onboard sensors and pre-set routes to travel in less busy 

areas. These shuttles typically travel at speeds of 25-35 mph and sometimes on fixed routes, so 

the technological advancement required is relatively small (May Mobility). Due to the lack of 

widespread V2X infrastructure in the United States, even in mid-sized cities, AV solutions are 

designed to operate independently, with remote monitoring as a backup (Zhang, 2025). We do 

see some interest in connectivity. A pilot connected intersection in Chicago will allow vehicles 

to receive signal phase timing information, which may be useful for future AV algorithms. But 

these initiatives are incremental. In short, mid-sized cities in the United States are testing the 

waters with standalone AV systems (from cars to shuttles), while Chinese cities are building 

integrated digital infrastructure from the ground up to support AVs. 

 

4.2.3 Urban Density and Development Impacts 

 

The deployment of AVs in intermediate density areas has not yet produced a clear effect, but 

the impact of density can be analyzed through early planning work. To enhance connectivity 

and controllable growth, China’s intermediate density cities integrate AV projects into their 

long-term urban development strategies. Chengdu and Suzhou treats AV industry to improve 



 

 29 

 

transportation connections instead of constructing another new road for human drivers, since 

their urban boundaries keep expanding. For example, some AV shuttles are deployed to the 

surrounding areas of Suzhou train station to connect the commercial and dwelling parcels. Land 

use may change as a result, making the supporting facilities in the area around the train station 

easy to reach without a private car. Similarly, Chengdu is integrating AVs into its high-tech 

development district. They have planned wide roads equipped with sensor systems, dedicated 

pickup/drop-off points for robotaxis, and mixed-use zoning that assumes on-demand ride 

services. These are evidence that AV mobility is being considered in the planning of new areas 

in China's medium-density cities. Planners expect self-driving vehicles to support the growth 

of mid-sized cities by connecting dispersed areas more efficiently and potentially reducing the 

need for private vehicles in new neighborhoods (Liu, 2024). 

 

On the U.S side, medium-sized cities have begun to incorporate AVinto their long-term 

mobility plans, but observable impacts are still limited to pilot areas. For example, although 

Pittsburgh's land use has not changed due to the presence of AV test vehicles and related 

businesses settled in the city, they have accumulated experience of the interaction between AV 

and pedestrians. Issues such as keeping safe bike lanes and preventing AVs from disrupting 

public transportation are of great concern to Pittsburgh residents (BikePGH, 2017). It has 

prompted local advocacy groups to require the deployment of AVs to be consistent with the 

concept of "people-oriented" street design, as described in NACTO's "Blueprint for 

Autonomous Urbanism". So, while nothing has changed on the ground yet, policy discussions 

in Pittsburgh are actively considering ways to ensure that AVs promote rather than harm urban 



 

 30 

 

livability (e.g., using AV data to optimize traffic signal timing, or requiring AVs to use caution 

in mixed traffic). So far, the direct impact of AV on Chicago's urban form has not been 

significant, but from a regional perspective, if autonomous trucks become popular, their impact 

can be imagined - logistics hubs may need to be redesigned to accommodate autonomous 

loading, and highways may also add dedicated autonomous lanes, thus changing infrastructure 

investment priorities. For passenger travel, if Chicago deploys autonomous shuttles connecting 

suburbs to transit hubs, it could attract more people living around, because people can take an 

autonomous shuttle for the last mile. Conversely, improved door-to-door autonomous ride 

services could encourage some people to live in more suburban or exurban areas, as such cars 

provide them with convenient access to work in distant city centers. One comparative study 

found that about 30% of American drivers might consider moving farther away if they had an 

autonomous car, compared with 42% in China (Guan, 2021). Urban sprawl caused by AVs may 

be a concern for both countries, but it may be more prominent in China's rapidly growing cities. 

In intermediate cities in the United States, where car dependence is already high, affordable 

autonomous commuting options could indeed expand the commuter base. As a result, some 

American urban planners warn that without policy intervention (such as congestion pricing or 

land use regulation), AVs may encourage suburban expansion—a continuation of the sprawl 

trend—rather than concentration. In summary, Chinese mid-sized cities are actively integrating 

AV into new urban developments, aiming to guide development and travel behavior to facilitate 

interconnected and efficient travel, while cities of similar size in the United States are still in 

the experimental stage, which may have an impact on spatial structure. 

 



 

 31 

 

4.3 Low-Density and Rural Settings 

 

Next, I compare AV programs in low-density and rural areas in China and the United States, 

focusing on how sparse environments affect deployment goals, feasibility, and urban 

development outcomes. The experience of small Chinese cities such as Yangquan with US 

examples such as suburban Phoenix and rural towns are compared. The economic 

considerations and social needs of low-density areas are very different from those of 

megacities—reducing the incentives for infrastructure-intensive vehicle-to-everything (V2X) 

systems while increasing the importance of vehicle-orientated AV technologies and specific 

use cases. 

 

4.3.1 Policy and Technological Adaption 

 

Yangquan, the smallest city in Shanxi Province, has begun adopting various autonomous 

driving services including robotaxi and self-driving buses despite their relatively low 

population density (People’s Daily Online, 2025). This reflects China’s top-down strategy to 

actively include smaller cities in national autonomous driving pilot projects. The national 

policy encourages the construction of intelligent transportation infrastructure across the country, 

so Yangquan and dozens of other cities have established autonomous driving test areas and 

upgraded smart roads with government support. In fact, even in areas with lower population 

density, local governments have worked with technology companies to deploy autonomous 

shuttles, Robotaxi and delivery robots in protected scenarios. Baidu has even deployed Apollo 



 

 32 

 

in Yangquan. These pilot projects often include building V2X on key sections of roads - for 

example, Yangquan’s network of connected roads can enable real-time traffic monitoring and 

management, thereby improving traffic safety and relieving traffic congestion. However, when 

constructing comprehensive intelligent infrastructure for AVs outside of major urban centers, 

it is unfavorable from an economic perspective. The sparse traffic and vast geographical area 

make expensive sensor networks and 5G coverage difficult to achieve (Ansarinejad, 2025). 

While China’s national strategy prioritizes the interconnection of densely populated cities, in 

areas with lower population density, it relies more on strong on-board autonomous driving 

technology, just like the US strategy. The Yangquan’s application model embodies a hybrid 

approach, with limited V2X infrastructure installed at strategic locations funded by the central 

government, while the AVs themselves are equipped to handle most of the perception and 

decision-making independently (CRSA, 2022). This vehicle-orientated adjustment is necessary 

because achieving truly ubiquitous "smart road" coverage in rural areas will be costly even in 

China. Nevertheless, government support ensures that smaller cities can experiment with AVs 

as a tool for local development and improved public services. Policy directives such as the 

2023 inter-ministerial notice on intelligent connected vehicles have also opened the door to 

L3/L4 autonomous driving pilots in many high- and low-density cities. In short, China's high-

level commitment has enabled the deployment of AVs even in areas with lower population 

density, although the scale of infrastructure investment will be adjusted according to local 

needs (McKinsey&Company, 2023).  

 

In the United States, AV deployment in low-density and rural areas is driven primarily by 



 

 33 

 

private sector initiatives and state-level policies. Unlike China’s top-down strategies, U.S.’s 

strategies are bottom-up and vehicle-focused. Instead of blanketing highways or suburbs with 

sensors and dedicated communications systems, U.S. companies are investing in developing 

advanced onboard Lidar, cameras, mapping, and AI to enable AVs to navigate existing roads 

independently. The vehicle-led approach to autonomous driving fits the U.S. situation, where 

AVs typically navigate highways and suburban streets with simple traffic conditions without 

the need for new roadside infrastructure. Waymo's robotaxi service in suburban Phoenix, for 

example, covers a wide swath of low-density neighborhoods without requiring special road 

modifications; the vehicles rely on their own suite of sensors to navigate wide arterial and dead-

end streets (Marshall, 2018). The Phoenix metropolitan area, known for its wide streets and 

pleasant weather, became an early AV hub precisely because its low-density suburban form 

reduces complications such as pedestrian traffic congestion (Marshall, 2018). In fact, Waymo 

has been expanding its operations in the Phoenix area to serve car-dependent communities, 

viewing suburban sprawl as an opportunity for AV ride-hailing services. Early deployments 

have focused on use cases where existing technology is feasible and where there is a clear 

private benefit, such as autonomous highway functionality and autonomous trucking between 

logistics hubs. Several companies are already piloting autonomous trucking on interstate 

highways in the Southwest, taking advantage of the fact that nearly half of U.S. truck miles are 

spread out on rural roads and interstates (Zarlf, Starks, Sussman & Kukreja, 2021). These trials 

use standard roads with minimal modification, highlighting a vehicle-orientated approach. In 

summary, the U.S. approach in low-density areas is characterized by the adaptive reuse of 

existing automotive infrastructure and a policy environment that allows autonomous innovators 



 

 34 

 

to operate on public roads with few special arrangements. The downside of this laissez-faire 

approach is that without public sector investment in connectivity, AV deployment in rural areas 

must overcome challenges such as spotty wireless network coverage and poorly maintained 

roads on their own. 

 

4.3.2 Impacts of Low Urban Density on AV deployment 

 

Next, I compare AV programs in low-density and rural areas in China and the United States, 

focusing on how sparse environments affect deployment goals, feasibility, and urban 

development outcomes. We contrast the experience of small Chinese cities such as Yangquan 

with US cases such as suburban Phoenix and rural towns. Unlike megacities, where economic 

considerations and social needs are different, low-density areas reduce the incentives for 

infrastructure-intensive vehicle-to-everything (V2X) systems, while increasing the importance 

of vehicle-focus AV technologies and specific use cases. 

 

Perhaps the most significant impact of low urban density on autonomous driving technology is 

the push-back against infrastructure-intensive solutions, which in turn will shift to vehicle-

focus autonomous driving. Chinese policymakers have proposed using V2X communications, 

smart traffic signals, and cloud-based coordination to manage the huge number of vehicles in 

densely populated mega-cities. But in rural areas or small towns, the return on such 

infrastructure is lower. Fewer vehicles and less traffic congestion make infrastructure 

economically and functionally unnecessary. Low-density environments are economically 



 

 35 

 

unable to build and maintain extensive intelligent road networks for a small number of vehicles. 

Researchers point out that remote locations make the development of fully connected "smart 

cities" more difficult, and sparsely populated areas often face connectivity barriers because 

cellular networks are unstable and far away, thus hindering V2X communications and real-time 

data sharing in rural areas (Ansarinejad, 2025). However, in the United States, this shift will 

not cause much trouble because from the beginning, its AV industry has assumed the hardware 

of the vehicle itself is the primary solution (AVIA, 2025). Considering that people cannot rely 

on external networked navigation on distant rural roads, automakers have equipped cars with 

high-tech sensors and high-definition maps helping them to interpret road conditions in real 

time. China's experience also reflects this trend. Cities like Yangquan do not yet have smart 

road infrastructure that supports the operation of AVs, and they still rely on their own hardware 

for daily operations. In low-density urban environments, both countries tend to prefer similar 

technical solutions. They both believe that highly AVs can cope with imperfect roads, 

unmarked country roads, or random hazards without continuous V2X input (USDOT, 2021). 

At the same time, safety and reliability technologies play an important role in future 

development. They need to be equipped with powerful sensors and reliable performance in all-

weather operation to operate safely in curving or twisting roads. In rural pilots, engineers found 

that vehicle sensors encountered challenges such as faded lane markings, gravel roads, and 

strong sunlight glare (Marketplace, 2024). Solving these problems with more advanced on-

board processing technology, such as improved vision algorithms or sensor fusion, instead of 

installing new road infrastructure on every section is more feasible from economic perspective. 

Therefore, from a qualitative perspective, low-density environments are more inclined to carry 



 

 36 

 

intelligent AVs. U.S. policymakers have supported this approach by withholding funding for 

roadside unit projects, while the Chinese government has admitted this by supporting projects 

that combine some infrastructure with strong vehicle AI in less populated areas, rather than just 

infrastructure. Despite differences in top-down and bottom-up philosophies between the two 

countries, both countries’ deployment of AVs outside of large cities will inevitably tend toward 

vehicle-focus autonomy. 

 

5. Summary Table 

 

Urban Density China United States 

Densely Populated Areas National strategies promote large-

scale pilot projects, Vehicle to 

Everything (V2X) + smart roads, 

AV as a supplement  to 

public transportation and reduce 

reliance on private cars 

Local decentralization (California 

allows testing, New York strictly 

restricts); Waymo/Cruise many 

problems in the experimental stage, 

no significant change in urban form 

Intermediate Density Areas National pilot extended to second-

tier cities (Suzhou/Chengdu); 

Smart infrastructure first and 

logistics scenarios first; New 

urban planning integrates AVs to 

optimize transportation 

connections 

Small-scale testing led by local 

governments (Pittsburgh/Chicago); 

independent vehicle-mounted system 

+ low-speed shuttle bus; concerns 

about conflicts between AVs and 

pedestrians, no significant changes in 

form. 

Low Density and Rural Areas National pilot projects in remote 

areas (Yangquan); selective V2X 

+ strong in-vehicle AI; improving 

traffic accessibility may aggravate 

the urban-rural gap 

Mainly led by private 

companies(Phoenix suburbs); 

completely dependent on vehicle 

sensors; suburban services expand 

commuter range, potential urban 

sprawl risk 
 

Table 6.1 summary table 

6. Discussion  

As the initially hypothesized, urban density was the primary variable influencing the 



 

 37 

 

adoption and impact of autonomous vehicles (AVs). However, further research revealed 

that, regardless of density, institutional structures and governance models play a more 

prominent role in shaping AV development than density itself. 

 

As the initially hypothesized, urban density was the primary variable influencing the 

adoption and impact of autonomous vehicles (AVs). However, further research revealed 

that, regardless of density, institutional structures and governance models play a more 

prominent role in shaping AV development than density itself. 

 

In China, a centralized planning system has enabled rapid deployment of AV technology 

through coordinated pilot zones, standardized technical frameworks, and direct 

infrastructure investment. Furthermore, China's AV deployment policies allow the 

government to shape industry development through public-private partnerships and 

subsidies. National support has enabled local governments to implement AV-related 

initiatives on a sustained and large scale, as demonstrated in medium-density cities such 

as Suzhou. In contrast, the United States, with its federal system, has decentralized AV 

regulation across state and municipal governments. Private companies in the United 

States often lead innovation in AV technology. This model, both directly and indirectly, 

has led to uneven AV deployment, with some cities experiencing rapid development 

due to lax state laws, while others face public resistance and regulatory ambiguity due 

to historical and social factors. 

 



 

 38 

 

 

7. Conclusion 

 

By a detailed comparative case studies of cities in China and the United States, I systematically 

analyze how urban density affects the deployment of AVs. With the application of SCOT theory 

and the Policy-Technology-Density framework, there are qualitative differences in AV industry 

development path between different urban densities. 

 

This study demonstrates that rather than a basic condition, urban density is actually a 

determinant of AV integration or deployment. The comparative case study between China and 

the United State shows that the density fundamentally shapes policy priorities and technology 

development direction. The complex transportation and spatial constraints of densely 

populated areas require strong technical and policy interventions. Chinese cities have adopted 

an infrastructure-supported approach, including extensive V2X networks and vehicle-road-

cloud systems to assist AVs to make up for the lack of single-vehicle capabilities, while US 

cities tend to rely on advanced on-board sensors and artificial intelligence to construct self-

sufficient AV systems in the absence of pervasive infrastructure. Despite their different 

governance actions, both countries demonstrate that high-density urban forms require 

proactive integration measures to ensure that AV enhance rather than disrupt urban mobility. 

Chinese intermediate cities have actively integrated AV pilot projects into new urban 

development plans to promote efficient and fit-in mobility gaps, while similar-sized US cities 

are still in the experimental stage of AV testing. Due to the lack of strategic planning, the 



 

 39 

 

latter may miss the opportunity to shape urban development in the new era. Therefore, the 

situation of intermediate cities dealing with further highlight the impact of density. Finally, in 

low-density and rural areas, the analysis of the two countries found convergence in approach. 

The sparse population and vast geographical area make large-scale infrastructure investment 

impractical and uneconomical, and both countries have shifted their focus to vehicle-focus 

solutions. In these areas, due to the limited returns of large-scale smart road systems, the 

deployment of AVs is usually with strong on-board autonomy. Overall, the main findings 

confirm that urban density significantly affects traffic patterns and deployment feasibility, 

determining which AV technologies and policies are most effective in each context. 

Policymakers may face unintended consequences when density is not properly considered. 

AVs may exacerbate suburban sprawl rather than improve urban efficiency if they are left to 

free market forces.   

However, urban density, as a single explanatory variable, has significant limitations in its 

ability to explain observed differences in AV development outcomes. Even in cities with 

similar population density parameters, such as New York City and Shanghai, the development 

trajectories and maturity of autonomous vehicles can exhibit significant divergence. These 

differences stem not only from the profound influence of urban spatial structural attributes 

such as density distribution and road network form, but also crucially from differences in 

governance models embedded in specific social systems, deeply embedded cultural contexts, 

and differentiated policy orientations and strategic intentions. Future research needs to adopt 

a multidimensional analytical model that encompasses not only density, policy instruments, 

and technology fit, but also the broader institutional context. 



 

 40 

 

 

8. Policy Recommendations 

 

Densely populated cities should leverage autonomous driving technology to enhance public 

transportation rather than simply make it become substitutes. Priority should be given to 

integrating AVs into high-density areas, enabling them to supplement public transportation and 

alleviate congestion in congested urban cores. There are three interconnected measures can 

optimize system efficiency: 1) Designated lanes with dynamic capacity allocation for 

autonomous shuttles during peak hours; 2) Adaptive traffic signals utilizing vehicle-to-

infrastructure communication to break the wall; 3) Digitally managed micro-zones for 

passenger transfers that minimize curbside conflicts. Investing in connected vehicle 

infrastructure, such as connected intersections and 5G corridors, can improve safety and 

navigation capabilities in these complex environments. At the same time, regulations should 

limit the unrestricted use of private AVs in densely populated downtown areas, such as through 

congestion pricing or additional charges for zero-occupancy vehicles. 

 

Intermediate cities have an opportunity to consider AVs as they shape growth patterns. Cities 

should carefully develop these technologies so that they promote sustainable development, 

rather than undermine it. They can encourage proactive planning and pilot projects to integrate 

AVs into evolving urban development and transportation networks. With growing populations, 

urban planners should incorporate AV-friendly designs into new districts and redevelopment 

projects, such as smart parking management and dedicated drop-off areas. Public-private 



 

 41 

 

partnerships can be established to test AV buses or shuttles as feeders to public transportation 

which improve “first-last-mile” connectivity. Policymakers should also use this modest-scale 

opportunity to guide travel behavior. Shared AV mobility services should be promoted rather 

than private AVs to prevent traffic congestion and suburban sprawl. Land use and zoning 

policies in these cities could be updated to support mixed-use, transit-oriented development 

models that leverage AVs for accessibility. 

 

The strategy for low-density areas is to rely on the technological capabilities of vehicles and 

selective public investment to achieve the benefits of AV access and safety improvements. They 

should focus on pragmatic deployment of AVs to address long-distance and sparse demand 

while avoiding expensive infrastructure investments. For suburban, exurban, and rural 

communities, the deployment of semi-AVs that can operate on traditional road networks is a 

practical solution. Policy frameworks should prioritize the expansion of shared AV services and 

address mobility issues for the elderly residents living in distant areas through targeted 

subsidies or rural transportation grants. Infrastructure transformation requires selective rather 

than comprehensive modernization, with upgrading cellular network connections on major 

routes and improving digital maps of key corridors to fully support navigation systems. At the 

same time, safety certification must continue to evolve to address challenges unique to rural 

areas, such as navigating unpaved roads and irregular traffic patterns. Land use planning needs 

to be proactively adjusted to prevent urban sprawl—even though the convenience of automated 

“door-to-door” travel may increase, zoning regulations need to be revised to maintain 

development density thresholds. This balanced approach combines technological adaptability 



 

 42 

 

with regulatory foresight to ensure that automated mobility enhances rather than destroys the 

fabric of low-density communities. 

 

 
  



 

 43 

 

Reference 

 
Ansarinejad, M., Ansarinejad, K., Lu, P., Huang, Y., & Tolliver, D. (2025). Autonomous 
vehicles in rural areas: A review of challenges, opportunities, and solutions. Applied Sciences, 
15(8), 4195. https://doi.org/10.3390/app15084195 
 
BBC. (2016, August 16). Self-driving cars: Who's responsible when things go wrong? 
https://www.bbc.com/news/technology-37117831 
 
Baidu robotaxis draw complaints from human drivers as service gains popularity. (2024, April). 
South China Morning Post. https://www.scmp.com/tech/big-tech/article/3269790/baidu-
robotaxis-draw-complaints-human-drivers-service-gains-popularity 
 
Bahrami, A. M. (2023). Infrastructure readiness for autonomous vehicles. Journal of Traffic 
and Transportation Management, 4(1), 1–4. https://doi.org/10.5383/JTTM.04.01.001 
 
Booth, L., Karl, C., Farrar, V., & Pettigrew, S. (2024). Assessing the impacts of autonomous 
vehicles on urban sprawl. Sustainability, 16(13), 5551. https://doi.org/10.3390/su16135551 
 
Brodsky, J. S. (2016). Autonomous vehicle regulation: How an uncertain legal landscape may 
hit the brakes on self-driving cars. Berkeley Technology Law Journal, 31(1), 851–878. 
https://doi.org/10.15779/Z38BG2H 
 
Chen, R., Gao, L., Liu, Y., Guan, Y. L., & Zhang, Y. (2023). Smart roads: Roadside perception, 
vehicle-road cooperation and business model. TechRxiv. 
https://doi.org/10.36227/techrxiv.24476473.v1 
 
Chicago Department of Transportation. (2021). Strategic Plan for Transportation. 
https://www.chicago.gov/content/dam/city/depts/cdot/CDOT%20Projects/Strategic_Plan/Stra
tegic_Plan_for_Transportation21.pdf 
 
China Daily. (2023, October 26). Pilot programs put more driverless vehicles on the road. 
https://www.chinadaily.com.cn/a/202310/26/WS653a745fa31090682a5eaf61.html 
 
Cugurullo, F., Acheampong, R. A., & Gueriau, M. (2021). The transition to autonomous 
vehicles: A socio-technical perspective. Cities, 110, 103064. 
https://doi.org/10.1016/j.cities.2020.103064 
 
Ding, F., Zhang, N., Li, S.-B., Bian, Y.-G., Tong, E., & Li, K.-Q. (2022). A survey of 
architecture and key technologies of intelligent connected vehicle-road-cloud cooperation 
system. Acta Automatica Sinica, 48(12), 2863–2885. https://doi.org/10.16383/j.aas.c211108 
 
European Environment Agency. (2025, January 13). Annex 2: Shared autonomous urban 



 

 44 

 

vehicles. In Transport and environment report 2022. 
https://www.eea.europa.eu/en/analysis/publications/transport-and-environment-report-
2022/annex-2-shared-autonomous-urban-vehicles 
 
Fagnant, D. J., & Kockelman, K. M. (2015). Preparing a nation for autonomous vehicles: 
opportunities, barriers and policy recommendations. Transportation Research Part A: Policy 
and Practice, 77, 167–181. https://doi.org/10.1016/j.tra.2015.04.003 
 
Fraedrich, E., Heinrichs, D., Bahamonde-Birke, F. J., & Cyganski, R. (2019). Autonomous 
driving, the built environment and policy implications. Transportation Research Part A: Policy 
and Practice, 122, 162–172. https://doi.org/10.1016/j.tra.2018.02.018 

 

Feng, C. (2024, July 10). Baidu robotaxis draw complaints from human drivers as service gains 

popularity. South China Morning Post. https://www.scmp.com/tech/big-tech/article/3269790/baidu-

robotaxis-draw-complaints-human-drivers-service-gains-popularity 

 
Gasgoo. (2024). Baidu’s six-generation driverless vehicle Apollo RT6 adopts LiDAR from 
ZVision. https://autonews.gasgoo.com/icv/70021750.html 
 
Guan, J., Zhang, S., D’Ambrosio, L. A., Zhang, K., & Coughlin, J. F. (2021). Potential impacts 
of autonomous vehicles on urban sprawl: A comparison of Chinese and US car-oriented adults. 
Sustainability, 13(14), 7632. https://doi.org/10.3390/su13147632 
 
Hasiri, A., & Kermanshah, A. (2024). Exploring the role of autonomous trucks in addressing 
challenges within the trucking industry: A comprehensive review. Systems, 12(9), 320. 
https://doi.org/10.3390/systems12090320 
 
Hawkins, A. J. (2024, December 10). Cruise's robotaxi service will shut down as GM pulls its 
funding. The Verge. https://www.theverge.com/2024/12/10/24318259/gm-cruise-shutdown-
robotaxi-super-cruise 
 
Heinrichs, D. (2016). Autonomous driving and urban land use. Retrieved from 
https://www.researchgate.net/publication/303480949_Autonomous_Driving_and_Urban_Lan
d_Use 
 
Heritage Foundation. (n.d.). Paving the way for autonomous vehicles: The future of mobility 
is here. https://www.heritage.org/government-regulation/report/paving-the-way-autonomous-
vehicles-the-future-mobility-here 
 
Kok, I. (2023). Planning and implementation of low- and zero-emission zones in cities. 
International Council on Clean Transportation. https://theicct.org/publication/planning-and-
implementation-lezs-zezs-in-cities-sept23/ 
 
Li, K., & Li, D. (2025). How autonomous vehicles shape urban traffic sustainability: An 



 

 45 

 

empirical study based on structural equation modeling. Sustainability, 17(6), 2589. 
https://doi.org/10.3390/su17062589 
 
Liu, C., Zyryanov, V., Topilin, I., Feofilova, A., & Shao, M. (2024). Investigating the impacts 
of autonomous vehicles on the efficiency of road network and traffic demand: A case study of 
Qingdao, China. Sensors, 24(16), 5110. https://doi.org/10.3390/s24165110 
 
Luo, Q., Saigal, R., Chen, Z., & Yin, Y. (2019). Accelerating the adoption of automated vehicles 
by subsidies: A dynamic games approach. Transportation Research Part B: Methodological, 
129, 226–243. https://doi.org/10.1016/j.trb.2019.09.011 
 
Marshall, A. (2018, December 8). 32 hours in Chandler, Arizona, the self-driving capital of the 
world. WIRED. https://www.wired.com/story/32-hours-chandler-arizona-self-driving-capital/ 
 
Mayor’s Office of New York City. (2024, March 27). Mayor Adams releases requirements, 
opens permit applications for responsible autonomous vehicle testing. 
https://www.nyc.gov/office-of-the-mayor/news/231-24/mayor-adams-releases-requirements-
opens-permit-applications-responsible-autonomous-vehicle 
 
Mayor’s Office of San Francisco. (2024, March). Mayor Lurie welcomes autonomous vehicles 
to Market Street as part of revitalization of downtown. https://www.sf.gov/news-mayor-lurie-
welcomes-autonomous-vehicles-to-market-street-as-part-of-revitalization-of-downtown-san-
francisco 
 
Makahleh, H. Y., Ferranti, E. J. S., & Dissanayake, D. (2024). Assessing the Role of 
Autonomous Vehicles in Urban Areas: A Systematic Review of Literature. Future 
Transportation, 4(2), 321–348. https://doi.org/10.3390/futuretransp4020017 
 
McAslan, D., Najar Arevalo, F., King, D. A., et al. (2021). Pilot project purgatory? Assessing 
automated vehicle pilot projects in U.S. cities. Humanities and Social Sciences 
Communications, 8, 325. https://doi.org/10.1057/s41599-021-01006-2 
 
Moore, M. A., Lavieri, P. S., Dias, F. F., & Bhat, C. R. (2020). On investigating the potential 
effects of private autonomous vehicle use on home/work relocations and commute times. 
Transportation Research Part C: Emerging Technologies, 110, 166–185. 
https://doi.org/10.1016/j.trc.2019.11.013 
 
Othman, K. (2022). Exploring the implications of autonomous vehicles: A comprehensive 
review. Innovative Infrastructure Solutions, 7, 165. https://doi.org/10.1007/s41062-022-00763-
6 
 
People's Daily Jiangsu. (2022, September 22). 苏州建设全国首个智能网联示范区 . 
http://js.people.com.cn/n2/2022/0922/c405572-40135522.html 
 



 

 46 

 

People’s Daily Online. (2023, November 2). Sichuan's first demonstration lines for commercial 
use of self-driving vehicles launched in Chengdu. https://en.people.cn/n3/2023/1102/c90000-
20092220.html 
 
Pennsylvania Autonomous Vehicle Policy Task Force. (2016, November 2). Pennsylvania 
autonomous vehicle testing policy: Final draft report. Pennsylvania Department of 
Transportation. https://www.pa.gov/content/dam/copapwp-
pagov/en/penndot/documents/research-planning-
innovation/researchandtesting/documents/av%20testing%20policy%20draft%20final%20rep
ort.pdf 
 
S&P Global. (2023). Mainland China autonomous vehicle development on a different track. 
https://www.spglobal.com/mobility/en/research-analysis/china-autonomous-vehicles-
development.html 
 
Schepis, D., Purchase, S., Olaru, D., Smith, B., & Ellis, N. (2023). How governments influence 
autonomous vehicle (AV) innovation. Transportation Research Part A: Policy and Practice, 178, 
103874. https://doi.org/10.1016/j.tra.2023.103874 
 
Science & Technology Daily. (2024, January). Development strategy of vehicle-energy-road-
cloud collaboration in the new technological situation. https://doi.org/10.15302/J-SSCAE-
2024.01.003 
 
Silva, Ó., Cordera, R., González-González, E., & Nogués, S. (2022). Environmental impacts 
of autonomous vehicles: A review of the scientific literature. Science of The Total Environment, 
830, 154615. https://doi.org/10.1016/j.scitotenv.2022.154615 
 
Smith, P. (2015, December 8). The man who tested the first driverless car in 1925 had a lot of 
guts. Jalopnik. https://www.jalopnik.com/the-man-who-tested-the-first-driverless-car-in-
1925-had-1792312207/ 
 
State of AV Industry Report. (2024). The AV Industry. 
https://theavindustry.org/resources/2024_StateOfAV.pdf 
 
Spencer, B. (2021, March 17). Baidu to develop China 5G driving project. ITS International. 
https://www.itsinternational.com/its5/its6/its7/news/baidu-develop-china-5g-driving-project 
 
Technology Review. Knight, W. (2016, September 13). Uber’s Pittsburgh project is a crucial 
test for self-driving cars. https://www.technologyreview.com/2016/09/13/7199/ubers-
pittsburgh-project-is-a-crucial-test-for-self-driving-cars/ 
 
Waltermann, J., & Henkel, S. (n.d.). Beyond acceptance: An innovative methodological 
approach to unveil societal perspectives on autonomous vehicles. SSRN. 
https://doi.org/10.2139/ssrn.5037485 



 

 47 

 

 
Wired. Marshall, A. (2024, February). General Motors cuts funding to Cruise robotaxis. 
https://www.wired.com/story/general-motors-cuts-funding-to-cruise-robotaxis/ 
 
Williams, J. (2023, October 16). The road to driverless cars: 1925–2025. Engineering.com. 
https://www.engineering.com/the-road-to-driverless-cars-1925-2025/ 
 
Xinhua News. (2024, March 17). China rolls out 'vehicle-road-cloud integration' trials across 
20 cities. https://english.news.cn/20240317/1c07d30eb3a0428da158fbd4dcd897b3/c.html 
 
Xinhua News Agency. (2024, March 20). Shanghai has over 2,000 km of roads allowing 
autonomous driving tests. Xinhua. 
https://english.news.cn/20240320/8b30f14c1caf47b88cf0bfa77b447de1/c.html 
 
Yicai Global. Zhang, Y. (2024, March). China rolls out ‘vehicle-road-cloud integration’ trials 
across 20 cities. https://www.yicaiglobal.com/news/china-rolls-out-vehicle-road-cloud-
integration-trials-across-20-cities 
 
Zarif, R., Starks, C., Sussman, A., & Kukreja, A. (2021, February 17). Autonomous trucks lead 
the way. Deloitte Insights. 
https://www2.deloitte.com/content/dam/insights/articles/us176547_cfs_fsi-
predictions_assisted-driving-technology-in-insurance/DI_FSI-Predictions_Assisted-
driving.pdf 
 
Zhang, X., Li, J., Zhou, J., et al. (2025). Vehicle-to-everything communication in intelligent 
connected vehicles: A survey and taxonomy. Automotive Innovation, 8, 13–45. 
https://doi.org/10.1007/s42154-024-00310-2