Cyberattacks are subtle and difficult to detect or differentiate from routine anomalies. Typically, manufacturing cybersecurity strategies rely on copies of network traffic that do not always indicate what is happening inside a system. And accessing such data from systems in near real time can jeopardize performance and safety.
To detect cyberattacks inside systems, researchers developed a framework that brings digital twin technology together with machine learning and human expertise to flag indicators of cyberattacks. When systems cannot be inspected while operating, digital twin technology often has provided data, like predicting when parts will require maintenance. In the case of an operational technology (OT) device such as a 3D printer used in this work, data flows to the digital twin, which artificial intelligence programs analyzed. Researchers then launched disturbances at the printer. Some were innocent anomalies; others were more nefarious.
The framework uses a process of elimination to distinguish cyberattacks. Artificial intelligence programs recognize the printer's normal and unusual conditions. They pass irregularities to other programs, which check them against a database. The system categorizes irregularities as anomalies or potential cyber threats. A human expert then confirms the system’s suspicions or inputs a new anomaly in the database. Theoretically over time, the system learns more, with experts inputting less.
Researchers checked the cybersecurity system’s results and found it correctly sorted cyberattacks and anomalies. Researchers plan to conduct more varied attacks, to ensure the framework's reliability and scalability.
National Association of Regulatory Utility Commissioners (NARUC) released its Considering Interoperability for Electric Vehicle Charging: A Commission Case Study, which provides findings from Connecticut’s pursuit of interoperability for its electric vehicle charging program. NIST supported the case study and Connecticut’s interoperability efforts.
After establishing an electric vehicle (EV) program, the Connecticut Public Utilities Regulatory Authority emphasized the need for an interoperable EV infrastructure. Its working group examined interoperability considerations, informed by NIST staff presentations. To inform other states, the study defines interoperability and its three levels, as found in NIST Framework and Roadmap for Smart Grid Interoperability Standards, Release 4.0.
The case study provides the Connecticut working group’s key considerations for EV charging infrastructure interoperability, which reflect those in NIST’s Smart Grid Framework 4.0.
Need for Use Cases: Interoperability needs vary with uses cases; entities will differ and so will their interfaces. Use cases help identify technologies and practices needed for an interoperable infrastructure.
Data and Communications: These are vital for a modern electricity grid and especially for an EV charging infrastructure connected to it.
Metrics for Meeting Standards: These metrics help assess a utility’s progress towards implementing and meeting standards.
Futureproofing: This means enabling interoperability that allows upgrades. When future iterations of technologies are available, smaller portions of systems can be upgraded without replacing them entirely.
Identification of Standards: These are needed to integrate EV charging infrastructure with the electric grid. This identification also involves assessing these standards’ testing and certification requirements.
The case study recommends determining “interoperability profiles” – a concept in NIST’s Smart Grid Framework 4.0. Standards often have options for integrating technologies into systems. Creating an interoperability profile focuses on one option (or a limited set of options) and prescribes a common integration approach.
NIST conducted the Named Data Networking (NDN) Community Meeting 2023 in Gaithersburg, MD at its National Cybersecurity Center of Excellence and online, on 2-3 March 2023. The meeting, organized by NIST’s Lotfi Benmohamed and external colleagues, is held annually and brings together researchers in the NDN community, which seeks an evolution from today's host-centric Internet architecture to a data-centric network architecture. This increased data-centric focus is seen as enhancing network usability and aiding mobile edge computing, Internet of Things, and augmented and virtual realities. Industry participation at this event included Airbus, Cisco Systems, Dell Technologies, Eluvio, Intel, and Operant Networks.
Researchers presented on a range of NDN-related issues, including:
Mission-Critical Emergency Operations: Researchers are pursuing disruption-tolerant networks, which have uses in these operations as well as in rural networking and manufacturing.
NDN for Vehicular Networking: Researchers reported that NDN is likely to improve interoperability between electronic control units and that a data-centric architecture has advantages for in-vehicle communications.
Secure Data Transfer for Mobile Health Infrastructure: Researchers developed an NDN architecture that addresses the security challenges related to smartwatches, fitness trackers, and body sensors.
NDN Opportunities in 5G/6G Core Networks: NIST researchers showed the benefits of NDN in 5G networks and how NDN could simplify a service-based architecture in 5G networks.
Panels also addressed the following:
Lessons Learned: Panelists discussed the lessons learned from more than 10 years of NDN research, and presented insights about specific aspects of the NDN architecture, including naming the data, which is central to NDN; different business cases for NDN deployment; and application programming interfaces for NDN application developers.
Named Data Metaverse: This panel discussed the use of the NDN technology for building Metaverse systems such as augmented/virtual realities for Internet-based experiences and existing gaps and related research opportunities to address Metaverse system requirements.
Time for Standardization? Despite considerable research, NDN has not significantly impacted the commercial space. This panel addressed whether standardization of NDN at this point in time is needed to achieve higher industry adoption of the technology.
International collaborators, including NIST, seeking to connect several 5G testbeds across the world, held a technical workshop at the University of Arizona in February 2023. The effort is called “Project Agility” and is pursuing testbeds that use 5G to access federated cloud computing, which ties together computing clouds in multiple sectors, to achieve faster and greater access to data from more sources.
The workshop builds on progress made in 2022 workshops held in Korea and at NIST in Gaithersburg, MD. The February 2023 workshop included participation by four NIST researchers. The workshop focused on designing a “network federation” across these testbeds – meaning a federation that intelligently configures a network for a requested service. This intelligent configuration will be enabled by artificial intelligence models in a knowledge plane, termed “Generic Autonomic Network Architecture (GANA),” across the network federation. NIST researchers will conduct a public working group to guide the development of this knowledge plane, or GANA.
International collaborators plan to provide cybersecurity for this knowledge plane, or GANA, using NIST’s “Zero Trust” principles for implementing an enterprise architecture, as described in NIST Special Publication 800-207 Zero Trust Architecture. Basically, Zero Trust means that every transaction on the network must be verified. It is a concept that is widely used beyond NIST.
International collaborators also planned the design of the “service federation” for these testbeds, which will enable the cloud federation to serve these testbeds. They will implement the service federation, in keeping with IEEE 2302-2021 Standard for Intercloud Interoperability and Federation. NIST researchers led the development of this standard, which is also based on Special Publication 500-332 The NIST Cloud Federation Reference Architecture.
NIST researchers Eugene Song and Kang Lee participated in the IEEE P2681 working group, which led to the development of a technical report, PES-TR102, MV Smart Grid Sensor and Sensor Systems: Measurement Accuracy and Uncertainty Considerations. The report was sponsored by the IEEE PES PSIM Sensors Task Force (TF) and posted in January 2023. The report supports the development of the proposed IEEE P2681 standard “Guide for Testing Medium Voltage (MV) Smart Grid Sensors and Intelligent Electronic Device (IED) Systems.”
NIST researchers contributed an architecture of smart sensor systems consisting of sensing, measurement, data processing, timing and synchronization, and network communication functions. These contributions helped model a generic smart sensor system, which was the outcome of NIST’s Smart Sensors project for the Smart Grids Testbed.
The report describes the challenges of implementing and deploying MV smart sensors and networks in smart grids. Notably, this report also introduces a conditional accuracy concept that addresses the overall measurement accuracy of smart sensor systems, within a specified range of operating conditions such as temperature, humidity, and range. The report provides examples of:
Medium voltage sensors within sensor networks with conditional accuracy
Smart sensor system applications in smart grids
Additionally, the report identifies and examines design, network, and external operational influences on the conditional accuracy of smart sensor systems. Lastly, end-user field testing of smart sensor systems is discussed, based on field operating conditions. More detail on measurement accuracy and interoperability testing will follow in IEEE P2681 standard, which currently is being developed.
NIST researchers developed an artificial intelligence protocol that can configure a test chamber to replicate the spatial characteristics of measured mmWave channels in industrial environments, thus allowing assessment of industrial wireless systems. Development of this protocol is described in “A Deep Reinforcement Learning Approach for Automated Chamber Configuration Replicating mmWave Directional Industrial Channel Behavior” which was presented by NIST researcher Mohamed Hany at the ARFTG-100th Microwave Measurement Symposium in January 2023.
The protocol meets a testing need for industrial wireless systems, which must be assessed for performance in industrial environments. Due to their highly reflective nature, industrial wireless channels differ from the characteristics of home and office channels. Thus, an industrial environment in which these wireless systems will operate must be replicated for industrial wireless equipment testing before deployment in “over-the-air” test chambers – which is often time-consuming.
NIST researchers developed a protocol that uses deep reinforcement learning – a subset of AI – that trains itself on a situation, getting rewards when it is right and costs when it is not. Researchers used the protocol to configure and tune the test chamber to the parameters needed for testing industrial wireless systems. Researchers then measured the performance of a specific mmWave system. The test method was validated by successfully comparing characteristics of channel measurements inside the test chamber, to measurements taken in a realistic environment.
This deep reinforcement learning protocol provides an automated approach to configuring a test chamber, thus allowing industrial wireless systems to be assessed for a wide range of environments, much faster and more efficiently than previously.
Complex systems consist of multiple networks – information, communications, power systems – and depend on each other to deliver services. Their interdependence also makes it possible for a local failure, or infection of a network, to spread to other systems. Thus, a major challenge is how to optimize investments to make complex systems robust enough to withstand such problems and keep operating with minimum or no disruptions – a subject that has been studied extensively.
Unlike past studies, NIST and University of Maryland researchers considered optimal investments for both resilience and recovery. In one approach, researchers devised an efficient gradient method to solve problems. To help determine optimum investments for a complex system’s resilience and recovery, researchers designed an algorithm which provided a quality solution, with bounds for optimum investment.
In another approach, researchers defined the technical conditions for a complex system’s operations and susceptibility to failures and infections and formulated a lower bound for an optimum investment, covering resilience and recovery. This formulation also helped determine a feasible solution and upward bound for an optimum investment. The effectiveness of both approaches was demonstrated, using numerical studies.
In February 2023, NIST’s Allison Barnard Feeney addressed several standards for capturing product manufacturing information in a presentation to the Secure Data Exchange/Interoperability workshop, which was held by the Department of Defense (DoD) Digital Manufacturing Enterprise Working Group. The workshop focused on developing a secure and interoperable manufacturing enterprise. This capability is seen as helping the DoD innovate its acquisition processes with the private sector.
Barnard Feeney covered where we are today regarding standards for product manufacturing information; this standardization is key to enabling systems to get the varying information needed for their respective functions in the manufacturing enterprise.
This portion of the presentation covered:
ISO 10303 application protocols: These are standards for software, enabling downstream use of product design data.
ISO TC 184/SC 4 standards: These standards apply to content, meaning, and quality of the digital data being exchanged.
Despite the need for more work on interoperable data flow, these standards have enabled accurate data exchange of geometry for repurposing data and long-term archiving. These standards also have been widely implemented.
Barnard Feeney also addressed the challenges that must be overcome to achieve standardized product information across the manufacturing enterprise. For example, initial Computer Aided Design output, a graphic representation, must be an output for a human readable medium. How to map semantic product information to geometric models has made considerable headway in the past decade, but requires a considerable amount of work to be able to exchange many other types of product manufacturing information.
Barnard Feeney ended by saying that ISO’s pursuit of “smart standards” promises to help overcome some of these challenges. Barnard Feeney said that the achievement of smart manufacturing could save industry $57.4 billion annually, citing NIST Government Contractor Report (GCR) 16-007, Economic Analysis of Technology Infrastructure Needs for Advanced Manufacturing.
Measurement science and standards are needed to support the safe and predictable operation of future automated vehicles (AVs). These vehicles have great potential to significantly impact our daily lives and improve the competitiveness of our economy.
NIST continues to build on a successful first-year Strategic and Emerging Research Initiatives (SERI) project, NIST and Autonomous Vehicles. (See 2022 NIST AV workshop and report.) NIST is supporting this expanded second-year SERI-funded, NIST Automated Vehicle Program to address system performance and measurement methods for AV sensors/perception, artificial intelligence (AI), cybersecurity, and AV communications. NIST is designing and establishing a system interaction testbed, which will involve both virtual and physical environments. This project brings together expertise from across NIST Laboratories and provides the basis for expanded stakeholder interactions, including a planned NIST AV workshop in September 2023.
Since 2019, NIST’s research on automated vehicles has been informed by its Automated Driving Systems Safety Measurement Technical Working Group, which meets regularly. Participants include representatives from over half of the automotive industry’s original equipment manufacturers; US Department of Transportation; several state transportation departments; and research universities such as Carnegie Mellon, University of California, University of Michigan, and Virginia Tech.
Today’s Efforts: The Technical Working Group seeks to determine what constitutes safe driving. This includes:
Determining ways of assessing whether automated vehicles meet the specifications of key driving behaviors or maneuver(s): Such behaviors include navigating an intersection, passing another vehicle, etc. These behaviors require specifications for their implementations and assessment. For example, staying in the center of a lane entails maintaining a speed within legal limits, and not deviating more than six inches from center line. Evaluating driving behaviors against their specifications will be used to determine and measure safe driving.
Assessing Human Trust in AI: Researchers in psychology and human factors assist this effort. Human trust is complex, evolving, and individual. Without passengers’ and vulnerable road user trust in automated vehicles, acceptance and adoption will be more difficult. A large part of the issue is replacing the human decision-maker with AI. This activity is developing a computer-digestible language for expressing human expectations of autonomous systems performance, such as those used to automate the driving task.
Those interested in joining the NIST Automated Driving Systems Safety Measurement Technical Working Group should contact Dr. Edward Griffor by email at edward.griffor@nist.gov.
