Applying ARPA-I: A Proven Model for Transportation Infrastructure

06.06.23 | 22分钟阅读 | Text byAndy GordonAleksandra SrdanovicCristina Velasquez

Executive Summary

2021年11月,国会通过了基础设施投资和就业法(IIJA), which included $550 billion in new funding for dozens of new programs across the U.S. Department of Transportation (USDOT). Alongside historic investments in America’s roads and bridges, the bill created the Advanced Research Projects Agency-Infrastructure (ARPA-I). Building on successful models like the Defense Advanced Research Projects Agency (DARPA) and the Advanced Research Program-Energy (ARPA-E), ARPA-I’s mission is to bring the nation’s most innovative technology solutions to bear on our most significant transportation infrastructure challenges.

ARPA-I must navigate America’s uniquely complex infrastructure landscape, characterized by limited federal research and development funding compared to other sectors, public sector ownership and stewardship, and highly fragmented and often overlapping ownership structures that include cities, counties, states, federal agencies, the private sector, and quasi-public agencies. Moreover, the new agency needs to integrate the strong culture, structures, and rigorous ideation process that ARPAs across government have honed since the 1950s. This report is a primer on how ARPA-I, and its stakeholders, can leverage this unique opportunity to drive real, sustainable, and lasting change in America’s transportation infrastructure.

How to Use This Report

这份报告强调了机会ARPA-I总统ents; orients those unfamiliar with the transportation infrastructure sector to the unique challenges it faces; provides a foundational understanding of the ARPA model and its early-stage program design; and empowers experts and stakeholders to get involved in program ideation. However, individual sections can be used as standalone tools depending on the reader’s prior knowledge of and intended involvement with ARPA-I.

An Opportunity for Transportation Infrastructure Innovation

2021年11月,国会通过了基础设施投资和就业法(IIJA) authorizing the U.S. Department of Transportation (USDOT) to create the Advanced Research Projects Agency-Infrastructure (ARPA-I), among other new programs. ARPA-I’s mission is to advance U.S. transportation infrastructure by developing innovative science and technology solutions that:


ARPA-I是一长串成功的ARPA的最新成员,它继续在国防,情报,能源和卫生部门提供突破性的创新。美国国防部建立了开拓者Defense Advanced Research Projects Agency(DARPA)1958年,为了响应苏联发射的卫星卫星,以开发和展示高风险,高回报的技术和能力,以确保美国军事技术优势和面对国家安全挑战。多年来,DARPA计划负责取得重大的技术进步,具有防御和国家安全以外的影响,例如互联网的早期阶段,全球定位系统(GPS)的创建以及对与之打击至关重要的mRNA疫苗的开发2019冠状病毒病。

鉴于通过DARPA计划获得的许多成功进步,政府复制了其他关键部门的ARPA模型,从而在国家情报部长办公室内进行了情报高级研究项目活动(IARPA),高级研究项目局 -能源部内的能源,以及最近的卫生与公共服务部内的高级研究项目机构 - 健康(ARPA-H)。

Now, there is the opportunity to bring that same spirit of untethered innovation to solve the most pressing transportation infrastructure challenges of our time. The United States has long faced a variety of transportation infrastructure-related challenges, due in part to low levels of federal research and development (R&D)spendingin this area; the fragmentation of roles across federal, state, and local government; risk-averse procurement practices; and sluggish commercial markets. These challenges include:

The Fiscal Year 2023 Omnibus Appropriations Bill awarded ARPA-I itsinitial appropriation在2023年初Net-Zero Game Changers Initiative。特别是,政府确定了智能移动性,清洁有效的运输系统,下一代基础设施构建,高级电力基础设施和清洁燃料基础设施为“净零游戏改变者”,ARPA-I可以在帮助发展方面发挥作用。


Unique Challenges of the Transportation Infrastructure Landscape


Lower Federal R&D Spending in Transportation Infrastructure

Federal R&D expenditures in transportation infrastructure lag behind those in other sectors. This gap is particularly acute because, unlike for some other sectors, federal transportation R&D expenditures often fund studies and systems used to make regulatory decisions rather than technological innovation. The table below compares actual federal R&D spending and sector expenditures for 2019 across defense, healthcare, energy, and transportation as a percentage of each sector’s GDP. The federal government spends orders of magnitude less on transportation than other sectors: energy R&D spending as a percentage of sector GDP is nearly 15 times higher than transportation, while health is 13 times higher and defense is nearly 38 times higher.

代理 ^1 Actual federal R&D spending, 2019 由行业添加的增值和2019年美国GDP的百分比 2019 federal R&D spending as % of sector GDP
Defense 546.9亿美元 $732 billion (3.4%) 7.5%
Health and Human Services $38.51 billion $1,452 billion (6.8%) ^2 2.7%
Energy 182.7亿美元 $607 billion (2.8%) ^3 3.0%
Transportation 11亿美元 6100亿美元(2.9%) ^4 0.2%


2. Excludes GDP value-adds relating to Social Assistance

3. Includes GDP value-adds relating to oil and gas extraction, utilities, and petroleum and coal products

4. Excludes GDP value adds relating to Warehousing

Public Sector Dominance Limits Innovation Investment

自1990年以来,美国研发的总投资增加了大约9次。在同一时期的研发投资来源时,私营和公共部门在1982年投资了大约相同数量的研发资金4 times greaterfor the private industry than the government.

While there are problems with the bulk of R&D coming from the private sector, such as innovations to promote long-term public goods being overlooked because of more lucrative market incentives, industries that receive considerable private R&D funding still see significant innovation breakthroughs. For example, the medical industry saw $161.8 billion in private R&D funding in 2020 compared to only $61.5 billion from federal funding. More than 75% of this private industry R&D occurred within the biopharmaceutical sector where corporations have profit incentives to be at the cutting edge of advancements in medicine.

The transportation sector has one robust domain for private R&D investment: vehicle and aircraft equipment manufacturing. In 2018, total private R&D was$52.6 billion。Private sector transportation R&D focuses on individual customers and end users, creating better vehicles, products, and efficiencies. The vast majority of that private sector R&D does not go toward infrastructure because the benefits are largely public rather than private. Put another way, the United States invests more than 50 times the amount of R&D into vehicles than the infrastructure systems upon which those vehicles operate.

Market Fragmentation across Levels of Government

Despite opportunities within the public-dominated transportation infrastructure system, market fragmentation is a persistent obstacle to rapid progress. Each level of government has different actors with different objectives and responsibilities. For instance, at the federal level, USDOT provides national-level guidance, policy, and funding for transportation across aviation, highway, rail, transit, ports, and maritime modes. Meanwhile, the states set goals, develop transportation plans and projects, and manage transportation networks like the interstate highway system. Metropolitan planning organizations take on some of the planning functions at the regional level, and local governments often maintain much of their infrastructure. There are also local individual agencies that operate facilities like airports, ports, or tollways organized at the state, regional, or local level. Programs that can use partnerships to cut across this tapestry of systems are essential to driving impact at scale.

当地机构的访问和能力有限,可以开发跨部门技术。他们可以访问有限的USDOT资金来进行试点技术,因此通常依靠市售技术来增加飞行员成功的可能性。当前过程的一个缺点是USDOT和基础架构所有者 - 操作员(IOOS)在开发创新技术方面发挥了更具被动的作用,而不是仅仅取决于仅部署市场就绪技术。

Multiple Modes, Customers, and Jurisdictions Create Difficulties in Efficiently Allocating R&D Resources




Having a matrixed understanding of the rapid technological evolution across transportation modes and their potential customers is critical to investing in and building infrastructure for the future, given that transportation infrastructure investments not only alter a region’s movement of people and goods but also fundamentally impact its development. ARPA-I is poised to shape learnings across and in partnership with USDOT’s modes and various offices to ensure the development and refinement of underlying technologies and approaches that serve the needs of the entire transportation system and users across all modes.




ARPAs should take risks.

An ARPA portfolio may be the closest thing to a venture capital portfolio in the federal government. They have a mandate to take big swings so should not be limited to projects that seem like safe bets. ARPAs will take on many projects throughout their existence, so they should balance quick wins with longer-term bets while embracing failure as a natural part of the process.

ARPAs should constantly evaluate and pivot when necessary.


ARPAs should stay above the political fray.




ARPA team members must move quickly and nimbly.

Trying to plan out the agency’s path for the next two years, five years, 10 years, or beyond is a futile effort and can be detrimental to progress. ARPAs require ultimate flexibility from day to day and year to year. Compared to other federal initiatives, ARPAs are far less bureaucratic by design, and forcing unnecessary planning and bureaucracy on the agency will slow progress.



Outcomes matter more than following a process.

ARPA PDs must be free to explore potential program and project ideas without any predetermination. The agency should support them in pursuing big and unconventional ideas unrestricted by a particular process. While there is a process to turn their most unconventional and groundbreaking ideas into funded and functional projects, transformational ideas are more important than the process itself during idea generation.


Things move quickly in an ARPA, and decisions must match that pace, so individuals such as fellows and PDs must work together to offer as much feedback as possible. Constructive pushback helps avoid blind alleys and thus makes programs stronger.


The ARPA Director sets the vision.

The Director’s vision helps attract the right talent and appropriate levels of ambition and focus areas while garnering support from key decision-makers and luminaries. This vision will dictate the types and qualities of PDs an ARPA will attract to execute within that vision.


Because the power of the agency lies within its people, ARPAs are typically flat organizations. An ARPA should seek to hire the best and most visionary thinkers and builders as PDs, enable them to determine and design good programs, and execute with limited hierarchical disruption. During this process, PDs should engage with decision-makers in the early stages of the program design to understand the needs and realities of implementers.

Contracting helps achieve goals.




职员transitions must be well facilitated to retain institutional knowledge.


Scaling should be built into the structure.



Similarly to the dedicated funding for scaling described above, technology-to-market advisors are responsible for thinking about how projects make it to the real world. They should work hand in hand with PDs even in the early stages of program development to provide perspectives on how projects might commercialize and become market-ready. Without this focus, technologies run the risk of dying on the vine—succeeding technically, but failing commercially.

A Primer on ARPA Ideation


Common Challenges

No clear diagnosis of the problem

Many challenges facing our transportation infrastructure system are not defined by a single problem; rather, they are a conglomeration of issues that simultaneously need addressing. An effective program will not only isolate a single problem to tackle, but it will approach it at a level where something can be done to solve it through root cause analysis.

Thinking small and narrow




Approaching solutions solely from a regulatory or policy angle


No explicit ARPA role

An ARPA should pursue opportunities to solve problems where, without its intervention, breakthroughs may not happen within a reasonable timeframe. If the public or private sector already has significant interest in solving a problem, and they are well on their way to developing a transformational solution in a few years or less, then ARPA funding and support might provide a higher value-add elsewhere.

Lack of throughline

ARPA程序考虑确定的问题tion should be present as themes throughout the opportunities chosen to solve them as well as how programs are ultimately structured. Otherwise, a program may lack a targeted approach to solving a particular challenge.


Human-centered design should be at the heart of how ARPA programs are scoped, especially when considering the scale at which designers need to think about how solving a problem will provide transformational change for everyday users.


Research programs should not be built with predetermined solutions in mind; they should be oriented around a specific problem to ensure that any solutions put forward are targeted and effective.



Keeping these common mistakes in mind throughout the design process ensures that programs are properly scoped, appropriately ambitious, and in line with the agency’s goals. With these guideposts in mind, idea generators should begin their program design in the form of a wireframe.


The first phase in ARPA program development is creating a program wireframe, which is an outline of a potential program that captures key components for consideration to assess the program’s fit and potential impact. The template below shows the components characteristic of a program wireframe.

Wireframe template

创建一个有血有肉的线框图,程序directors work backward by first envisioning a future state that would be truly transformational for society and across sectors if it were to be realized. Then, they identify a clearly-articulated problem that needs solving and is hindering progress toward this transformational future state. During this process, PDs need to conduct extensive root cause analysis to consider whether the problem they’ve identified is exacerbated by policy, regulatory, or environmental complications—as opposed to those that technology can already solve. This will inform whether a problem is something that ARPA-I has the opportunity to impact fundamentally.

Next, program directors identify a promising opportunity—such as a method, approach, or technology—that, if developed, scaled, and implemented, would solve the problem they articulated and help achieve their proposed future state. When considering a promising opportunity, PDs must assess whether it front-runs other potential technologies that would also need developing to support it and whether it is feasible to achieve concrete results within three to five years and with an average program budget. Additionally, it is useful to think about whether an opportunity considered for program development is part of a larger cohort of potential programs that lie within an ARPA-I focus area that could all be run in parallel.

Most importantly, before diving into how to solve the problem, PDs need to articulate what has prevented this opportunity from already being solved, scaled, and implemented, and what explicit role or need there is for a federal R&D agency to step in and lead the development of technologies, methods, or approaches to incentivize private sector deployment and scaling. For example, if the private sector is already incentivized to, and capable of, taking the lead on developing a particular technology and it will achieve market readiness within a few years, then there is less justification for an ARPA intervention in that particular case. On the other hand, the prescribed solution to the identified problem may be so nascent that what is needed is more early-stage foundational R&D, in which case an ARPA program would not be a good fit. This area should be reserved as the domain of more fundamental science-based federal R&D agencies and offices.

DARPA投资在mRNA中进行了说明。尽管美国国立卫生研究院为初步基础研究做出了重大贡献,但DARPA认识到能够快速扩展和制造治疗疗法的技术差距,促使该机构启动自动诊断以实现预防和治疗方法(Adept),以开发技术以响应技术来响应技术以响应技术来响应技术以做出反应传染病威胁。通过Adept,DARPA在2011年通过2500万美元研究和开发其Messenger RNA Therapeutics平台。九年后,Moderna成为继Pfizer-Biontech接收的第二家公司Emergency Use Authorizationfor its COVID-19 vaccine.

Another example is DARPA’s role in开发互联网as we know it, which was not originally about realizing the unprecedented concept of a ubiquitous, global communications network. What began as researching technologies for interlinking packet networks led to the development of ARPANET, a pioneering network for sharing information among geographically separated computers. DARPA then contracted BBN Technologies to build the first routers before becoming operational in 1969. This research laid the foundation for the internet. The commercial sector has since adopted ARPANET’s groundbreaking results and used them to revolutionize communication and information sharing across the globe.

Wireframe Refinement and Iteration

To guide program directors through successful program development, George H. Heilmeier, who served as the director of DARPA from 1975 to 1977, used to require that all PDs answer the following questions, known as theHeilmeier Catechism,作为他们为新计划的推销的一部分。这些问题应用于完善线框并设想该程序的外观。特别是,线框改进应在扩展到其余问题之前检查前三个问题。

除了Heilmeier教义问答,一系列的屁股essments and lines of questioning should be completed to pressure test and iterate once the wireframe has been drafted. This refinement process is not one-size-fits-all but consistently grounded in research, discussions with experts, and constant questioning to ensure program fit. The objective is to thoroughly analyze whether the problem we are seeking to solve is the right one and whether the full space of opportunities around that problem is ripe for ARPA intervention.


Once a high-impact program has been designed, the next step is to rigorously pressure test and develop a program until it resembles an executable ARPA program.

Applying ARPA Frameworks to Transportation Infrastructure Challenges

通过使用此框架,可以评估运输基础架构中的任何问题或机会,以作为ARPA级别的想法。专家和利益相关者的想法生成对于创建ARPA-I程序的有效投资组合至关重要,因此Idea Generator必须使用此框架和一组定义的焦点领域来开发有希望的计划线框。ARPA-I的初始焦点领域包括安全性,气候和弹性以及数字化,并具有公平性和可访问性,作为每个焦点区域内的基本考虑。


Note: The following wireframes are samples intended to illustrate ARPA ideation and the wireframing process, and do not represent potential research programs or topics under consideration by the U.S. Department of Transportation.

Next-Generation Resilient Infrastructure Management

A Digital Inventory of Physical Infrastructure and Its Uses

线框开发Next Steps


Wireframes are intended to be a summary communicative of a larger plan to follow. After further iteration and exploration of the factors outlined above, what was first just a raw program wireframe should develop into more detailed documents. These should include an incisive diagnosis of the problem and evidence and citations validating opportunities to solve it. Together, these components should lead to a plausible program objective as an outcome.




为未来的更新注册事件,召开s, and other opportunities for you to work in support of ARPA-I programs and partners,点击这里