Cyber-physical systems (CPS) are “engineered systems that are built from, and depend upon, the seamless integration of computational algorithms and physical components.”1 CPS can be small and closed, such as an artificial pancreas, or very large, complex, and interconnected, such as a regional energy grid. CPS engineering2 focuses on managing interdependencies and impact of physical aspects on cyber aspects, and vice versa. With the development of low-cost sensing, powerful embedded system hardware, and widely deployed communication networks, the reliance on CPS for system functionality has dramatically increased. These technical developments in combination with the creation of a workforce skilled in engineering CPS will allow the deployment of increasingly capable, adaptable, and trustworthy systems.
CPS ENGINEERING AND THE CPS WORKFORCE
CPS are already widely deployed and used today. Examples include automobiles that sense impending crashes and perform various tasks to
1 Definition from National Science Foundation, 2016, “Cyber-Physical Systems,” Program Solicitation 16-549, NSF document number nsf16549, March 4, https://www.nsf.gov/publications/pub_summ.jsp?ods_key=nsf16549.
2 The committee uses the terms “CPS engineering” and “CPS engineer” to mean a set of skills and knowledge needed to design and build a CPS and a person with those skills; the terms are not limited to a set of credentials or to someone who has a degree or certification in CPS.
protect passengers and medical devices that sense glucose levels or the heart’s rhythm and intervene to restore normal body function. As these examples illustrate, CPS often support critical missions that have significant economic and societal importance and raise significant safety and cybersecurity concerns. However, today’s practice of CPS system design and implementation is often ad hoc, not taking advantage of even the limited theory that exists today, and unable to support the level of complexity, scalability, security, safety, interoperability, and flexible design and operation that will be required to meet future needs.
Engineers responsible for developing CPS but lacking the appropriate education or training may not fully understand at an appropriate depth, on the one hand, the technical issues associated with the CPS software and hardware or, on the other hand, techniques for physical system modeling, energy and power, actuation, signal processing, and control. In addition, these engineers may be designing and implementing life-critical systems without appropriate formal training in CPS methods needed for verification and to assure safety, reliability, and security.
A workforce with the appropriate education, training, and skills will be better positioned to create and manage the next generation of CPS solutions. Building this workforce will require attention to educating the future workforce with all the required skills—integrated from the ground up—as well as providing the existing workforce with the needed supplementary education.
It proved difficult to obtain comprehensive data on industrial demand for CPS skills, and the committee was not in a position to commission systematic surveys to collect such information itself. Instead, the committee has relied on the perspectives of industry experts, including those who briefed the committee or who participated in the two workshops convened during its study. It was also apparent from these presentations that the CPS field will continue to evolve as new applications emerge and as more research is done.
FINDING 1.1: CPS are emerging as an area of engineering with significant economic and societal implications. Major industrial sectors such as transportation, medicine, energy, defense, and information technology increasingly need a workforce capable of designing and engineering products and services that intimately combine cyber elements (computing hardware and software) and physical components and manage their interactions and impact on the physical environment. Although it is difficult to quantify the demand, a likely implication is that more CPS-capable engineers will be needed.
FINDING 1.2: The future CPS workforce is likely to include a combination of (1) engineers trained in foundational fields (such as electrical and computing engineering, mechanical engineering, systems engineering, and computer science); (2) engineers trained in specific applied engineering fields (such as aerospace and civil engineering); and (3) CPS engineers, who focus on the knowledge and skills spanning cyber technology and physical systems that operate in the physical world.
The mix of programs offered by universities will reflect the perspectives of individual institutions, their resources, and the demand universities see from students and their employers, and in turn affect the educational backgrounds of the CPS workforce. Over time, as the field itself changes and matures, education and employer demand will co-evolve.
FINDING 1.3: Given that most entry-level engineering and computer science positions are filled by undergraduates, it is important to incorporate CPS into the undergraduate engineering and computer science curricula.
RECOMMENDATION 1.1: The National Science Foundation, together with universities, should support the creation and evolution of undergraduate education courses, programs, and pathways so that engineering and computer science graduates have more opportunities to gain the knowledge and skills required to engineer cyber-physical systems. The efforts should be complemented by initiatives to augment the skills of the existing workforce through continuing education and master’s degree programs.
CPS PRINCIPLES, FOUNDATIONS, SYSTEM CHARACTERISTICS, AND COMPLEMENTARY SKILLS
This section summarizes the knowledge and skills needed to engineer CPS. It is derived from an examination of existing courses, programs, and instructional materials as well as consideration of the topics highlighted in comments from industry experts. The emphasis is deliberately on core principles and foundations reflecting the challenge of packing the material needed to span both cyber and physical aspects into an already crowded engineering curricula.
The committee has identified the following four broad areas for CPS education programs to cover:
- Principles that define the integration of physical and cyber aspects in such areas as communication and networking, real-time operation, distributed and embedded systems, physical properties of hardware and the environment, and human interaction.
- Foundations of CPS in (1) basic computing concepts, (2) computing for the physical world, (3) discrete and continuous mathematics, (4) crosscutting applications, (5) modeling, and (6) system development.
- System characteristics required of CPS, such as security and privacy; interoperability; reliability and dependability; power and energy management; safety; stability and performance of dynamic and stochastic systems; and human factors and usability.
Each area is briefly outlined in the sections below (and discussed in more detail in Chapter 2).
CPS bridges engineering and physical world applications and the computer engineering hardware and computer science cyber worlds. Basic principles of the physical world include physics, mathematical modeling, analysis, and algorithm and systems design and deal with their associated uncertainty and risk. Principles of the computer engineering and computer science (cyber) worlds deal with embedded computation and communications hardware systems, software programming, and networking, Because sensors are a key hardware bridge between the physical and cyber worlds, it is important to understand the properties of sensors and their real-world behavior, and techniques for processing the signals they produce. Control theory is an important tenet of CPS; relevant elements include stability, optimization, and how to control distributed, digital systems.
Foundations of CPS
Drawing on these principles, the committee identified the following six key overarching intellectual foundations for a CPS curriculum:
- Basic computing concepts beyond those covered in a couple of introductory programming courses, such as embedded hardware, data structures, automata theory, and software engineering.
- Computing for the physical world, which involves understanding physical world properties, real-time embedded systems, and computing resource constraints such as power and memory size.
- Discrete and continuous mathematics beyond calculus, such as differential equations, probability and stochastic processes, and linear algebra.
- Cross-cutting application of sensing, actuation, control, communication, and computing reflecting the central role of interactions between physical and cyber aspects and the reliance on control over communication networks, sensing, signal processing, and actuation with real-time constraints.
- Modeling of heterogeneous and dynamic systems integrating control, computing, and communication, with emphasis on uncertainty and system heterogeneity, including such techniques as linear and nonlinear models, stochastic models, discrete-event and hybrid models, and associated design methodologies based on optimization, probability theory, and dynamic programming.
- CPS system development, especially for safety-critical, high-confidence, and resilient systems, requires a life-cycle view from initial requirements to testing to certification and in-service use, including formal verification and validation procedures and adaptable designs that can accommodate system evolution.
FINDING 2.1: Core CPS knowledge involves not only an understanding of the basics of physical engineering and cyber design and implementation, but understanding how the physical and cyber aspects influence and affect each other.
RECOMMENDATION 2.1: Cyber-physical systems educational programs should provide a foundation that highlights the interaction of cyber and physical aspects of systems. Most current courses fail to emphasize the interaction, implying that new courses and instructional materials are needed.
Many CPS are large, complex, and/or safety critical. Successful development of such systems requires knowledge of how to ensure that systems possess the following characteristics:
- Security and privacy,
- Reliability and dependability,
- Power and energy management,
- Stability and performance, and
- Human factors and usability.
These topics are best introduced early and infused throughout the CPS curriculum in coursework and projects, much as the best practice in engineering is to address these issues from the outset of system design.
The growing scale and complexity of engineering systems mean that engineers are increasingly working collaboratively with experts from multiple disciplines. “Soft” skills—in such areas as communication, flexibility, and an ability to work on teams, including multiple disciplines—are of particular importance for CPS engineering because the work is inherently interdisciplinary. The pace of change in science and engineering knowledge generally and the newness and rapid flux of CPS suggest that CPS courses and programs emphasizing learning and critical thinking, as well as specific techniques and methods, are needed.
PATHS TO CPS KNOWLEDGE
There will be multiple paths for attaining CPS knowledge and skills (Box S.1). One reason is that the workforce is likely to include both domain experts who are knowledgeable about CPS principles and a new type of engineer who is an expert at the intersection of cyber and physical issues. Another reason is that many different approaches will be undertaken at colleges and universities depending on their present circumstances, such as existing department structures and curricula, faculty expertise, and available resources.
Designing a CPS degree is quite complex and involves, for example, a careful balancing of physical and cyber aspects and general CPS and application knowledge. Because CPS degree curricula are in their infancy, they will doubtless evolve substantially as CPS are more widely deployed. Moreover, CPS programs will doubtless share with most engi-
neering degree programs the challenge of prioritizing topics to fit in a manageable 4-year program of study.
FINDING 3.1: The diversity of current departmental structures, faculty expertise and interests, and curricula suggest that there are multiple feasible and appropriate models for strengthening CPS engineering. The committee envisions that universities will (1) enrich current engineering programs with CPS content, (2) create CPS survey courses, (3) create new master’s-level CPS degrees, and, ultimately, (4) develop new undergraduate CPS engineering degree programs.
Many universities may not currently have the expertise or resources to establish extensive CPS education programs. A useful alternative in these cases would be to forge more limited partnerships among several departments to implement jointly taught courses. For example, key CPS content could be introduced into mechatronics, robotics, or transportation courses. Doing so over time could help reduce the burdens associated with infusing CPS throughout engineering and building the courses needed to implement a CPS program.
FINDING 3.2: Because CPS engineering centers on the interaction of physical and cyber aspects of systems, it will often not be sufficient to create CPS curricula by simply combining material from existing courses. New courses will need to be designed.
RECOMMENDATION 3.1: The National Science Foundation should support the development of university education programs that define a path and plan for the creation of a cyber-physical systems engineering degree.
RECOMMENDATION 3.2: The National Science Foundation, professional societies, and university administrations should support and consider allocating resources for the development of new cyber-physical systems (CPS)-focused courses within existing engineering programs, new CPS-specific classes for CPS engineering majors and minors, and an overall curriculum for an undergraduate CPS engineering degree program.
RECOMMENDATION 3.3: Universities should consider adding cyber-physical systems content to freshman-level introductory courses for students in all areas of engineering and computer science.
RECOMMENDATION 3.4: Engineering schools, by-and-large, have already redesigned their curricula to emphasize project-based learning. Because this is especially important for cyber-physical systems (CPS) education, these project-based courses should be extended to support CPS principles and foundations.
OPPORTUNITIES AND OBSTACLES FOR INSTITUTIONALIZING CPS CURRICULA
Several obstacles stand in the way of building successful CPS programs. The nature of CPS makes it difficult to develop and teach CPS-focused curricula. Moreover, although students may be interested in CPS technologies or in the applications that CPS enable, they may not realize that they ought to seek out courses or a program that emphasizes CPS knowledge and skills. Also, few mechanisms exist to support extensive faculty commitment to a new interdisciplinary discipline, which makes it hard to develop, recruit, or retain the faculty needed to provide an up-to-date CPS education for undergraduate students. Moreover, an array of resources—from new textbooks to laboratory equipment—is needed to support any new curriculum.
Drawing Students to CPS
At the undergraduate level, one key will be exposing STEM-oriented students to the existence of the field of CPS, its links to related areas like robotics and the Internet of Things (IoT), and the potential benefits of its formal study. One important opportunity is to include CPS as part of freshman “introduction to engineering” programs across engineering and not just in computer science and electrical engineering.
FINDING 4.1: Although there are many STEM courses and programs at the high school and undergraduate level that introduce the students to some CPS elements, such programs often do not provide a broad introduction to CPS foundations and principles and tend to be overly focused either on simplistic applications or discipline-centric content.
RECOMMENDATION 4.1: Those developing K-12 science, technology, engineering, and mathematics (STEM) programs and educating and training STEM teachers should consider opportunities to enrich these programs with cyber-physical systems (CPS) concepts and applications in order to lay intellectual foundations for future work and expose students to CPS career opportunities.
FINDING 4.2: Incoming college students appear to be unfamiliar with the term CPS, CPS concepts, and job opportunities in CPS. They are, however, drawn to courses and programs in more widely visible CPS-related topics such as robotics, the Internet of Things (IoT), health care, smart cities, and the Industrial Internet.
RECOMMENDATION 4.2: Those developing cyber-physical systems engineering courses and programs should consider leveraging the visibility of and student interest in areas such as robotics, the Internet of Things, health care, smart cities, and the Industrial Internet in descriptions of careers, courses, and programs and when selecting applications used in courses and projects.
Recruiting, Retaining, and Developing the Needed Faculty
Faculty teaching CPS courses will be most effective if they are able to draw on expertise in particular aspects of CPS, knowledge of the other aspects of a complete CPS system, and domain- or application-specific needs. Today, most CPS education (and research) is being performed by a small number of faculty members who previously established themselves in a related field and then ventured into this newer, more interdisciplinary field.
In the long term, academic institutions will have opportunities to recruit new faculty who have graduated with a CPS degree or specialization and who have a record of conducting CPS-specific research as well as people with industrial experience with CPS engineering. Indeed, some institutions have already begun explicitly looking for such individuals. Both research funding and opportunities for academic advancement are needed to develop a pool of faculty. The National Science Foundation’s Cyber-Physical Systems program has helped build an academic community around CPS and foster links between academia and industry. The parallel development of several well-recognized CPS conferences and the creation of a new CPS journal have also made it easier for faculty with a multidisciplinary profile to establish themselves as CPS researchers and still meet the academic evaluation criteria. Nevertheless, it will take time and investment to build the necessary complement of faculty to educate those who engineer CPS.
FINDING 4.3: Because CPS is a new field that draws on multiple disciplines, not all institutions can be expected to have enough faculty with the requisite knowledge to teach all of the courses needed for a CPS degree program.
RECOMMENDATION 4.3: The National Science Foundation should support the development of cyber-physical systems faculty through the use of teaching grants and fellowships.
Despite the challenges of entering a new field, young faculty may have an advantage becoming leaders in the CPS field, given the novelty of the area, because they do not need to compete with the large number of well-established and well-recognized leaders found in more mature fields.
Developing Needed Courses and Instructional Materials
Although the committee was encouraged by the release of several textbooks during the course of its work, the number of textbooks, curricular materials, and laboratory facilities that exist to support CPS remains limited. Just as merely regrouping current classes will not yield a CPS curriculum, current texts may not fully incorporate the effects of the physical system on cyber technology, and vice versa. Furthermore, often the complexity of CPS demands that students gain a full understanding of how the physical environment impacts these systems. Realistic models can provide some of this knowledge, but testbeds will be needed for students to fully realize the constraints the physical environment can create. These testbeds are expensive to create and maintain, and many universities do not have, or will not allocate, the resources to create such testbeds.
FINDING 4.4: If they are to teach new CPS courses and build CPS programs, universities will need to allocate time and resources to develop CPS course materials and to provide the necessary laboratory space and equipment (including both virtual and physical testbeds).
FINDING 4.5: Testbeds are needed to provide students with sufficiently realistic applications and problems. These can be both virtual and physical and can be remotely accessed and shared among multiple institutions and developed and operated in cooperation with industry.
RECOMMENDATION 4.4: The National Science Foundation, professional societies, and universities should support the development and evolution of cyber-physical systems textbooks, class modules (including laboratory modules), and testbeds. These parties should partner with industry in developing and maintaining realistic testbeds.
* * *
As CPS become more pervasive, demand will grow for a workforce with the capacity and capability to design, develop, and maintain them. An understanding of not only the cyber or the physical aspects of systems, but also their interactions will become more and more valuable. A workforce with these skills will be better positioned to help industry pursue current and future advances across the myriad applications for CPS. The actions recommended in this report point to ways to ensure that aspiring engineers and computer scientists are equipped with the skills necessary to meet the demand for a modern CPS workforce.