Download PDF Spring Bridge on Postpandemic Engineering March 14, 2021 Volume 51 Issue 1 This issue is dedicated to the future of manufacturing. A stellar slate of experts present diverse experiences and perspectives from industry, a national laboratory, and academia. Together the articles provide informative coverage and holistic views on the future of advanced manufacturing, leveraging new and emerging technologies, desired infrastructure, innovative approaches, and a resilient supply chain to fortify US manufacturing competitiveness in the coming years. Next-Generation IIoT: A Convergence of Technology Revolutions Thursday, April 1, 2021 Author: Barbara L. Goldstein and Kate A. Remley Through measurements and standards, NIST aims to facilitate the postpandemic adoption of efficient, secure, and decentralized technology for the industrial sector. The pandemic forced many on a personal journey of digital transformation akin to that required of industry and much of the workforce. People had to come to terms with the fact that the usual ways of doing business could not simply be continued. From Digitization to Digitalization to Digital Transformation The shift from an office to a home environment required successful digitization—having automation tools in place and the necessary information in a digital format compatible with those tools. This transformation is parallel to manufacturers reaping the benefits of the Third Industrial Revolution, when manual tasks were automated with computers and robotics. After digital readiness comes connection—in the home and to the broader world—and the ability to transmit and interpret data as they flow from system to system, location to location. This is comparable to progressing from the third to the Fourth Industrial Revolution, which added connectivity to the robotics and infrastructure of a stand-alone factory. Just as individuals are relearning how to work in a geographically dispersed but interconnected world, in this era of digitalization manufacturers are reshaping their business models to automate supply and logistics by directly linking inventory and logistics systems across the supply chain, to automate equipment repair and maintenance through interconnected vendor/manufacturer systems, and to gather information directly from products in the field to adapt and accelerate the development of the next generation of offerings. The next step, digital transformation, the hallmark of Industry 4.0, has no playbook. It is a system-level restructuring that evolves once machine-to-machine automation can be trusted not only to inform workers but to run operations. Postpandemic manufacturing will likely build on these goals by capitalizing on decentralized systems that allow flexible and reconfigurable factory floors, and by leveraging low-cost yet accurate internet of things (IoT)–based technologies to maximize productivity and efficiency. All are important for economic recovery and will allow small and midsize factories to better compete in the global marketplace. Here we describe some of the research underway at the National Institute of Standards and Technology (NIST) to support manufacturing in the postpandemic world. The Factory of the Future The pandemic made it even more urgent that manufacturers accelerate their journey from digitization to full digital transformation—not just adopting automation tools but learning how to leverage them into new ways of doing business. Over the past decade or so, the manufacturing sector has embraced the use of wireless industrial internet of things (IIoT) technologies for factory enterprise applications, such as asset and supply chain management. Moving forward, particularly in the postpandemic world, the manufacturing sector is looking to decentralize decisions, implement remote monitoring of physical processes, and enable manufacturing tools to communicate and cooperate with each other in real time. High reliability will be critical in such machine-to-machine (M2M) real-time manufacturing processes. Low-cost, low-profile sensors and actuators operating accurately at high speeds will allow decentralization as machines talk to each other, monitoring their status and taking themselves offline before damage can occur. The development of reliable wireless M2M communications in a highly reflective, dynamically changing factory floor environment is also a key enabler of IIoT, motivating system designers to investigate new wireless technologies such as adaptive millimeter-wave (mmWave) networks that can reconfigure on the fly and reliably transfer large amounts of data over wide bandwidths. Testing and verifying the performance of adaptive networks in an over-the-air (OTA) condition is not a trivial undertaking and is the focus of much research at NIST. As humans are removed from the decision--making loop, it is essential that machines draw on accurate and always available information. Trends in Sensing Sensors bridge the physical and digital worlds, feeding information about the former to the automation tools that industry increasingly trusts to not only monitor and diagnose but predict and act—at speeds that don’t allow for human oversight. Sensors must be reliable reporters of conditions on the ground, so that digital twins are faithful mimics of a factory environment, machine-driven controls respond to real conditions, and equipment self-diagnoses are based on accurate information. NIST is addressing this need through the NIST on a Chip (NOAC) program, which is developing quantum-based sensors that are fit-to-function and reliable in the field without needing calibration, and can be trusted to give either the right answer or no answer at all. The sensors being developed in the NOAC program draw their intrinsic accuracy from fundamental properties of nature, such as the fact that a cesium atom can be counted on to always vibrate at a known -frequency—otherwise it wouldn’t be cesium. In fact, the entire International System of Units (SI, or metric system) was redefined in 2019 to be based only on such fundamental properties, making obsolete the last physical artifact, a platinum-iridium cylinder (stored in a vault outside Paris) that was by definition equal to 1 kilogram no matter what it really weighed. This recent sweeping redefinition of the metric system created new opportunities for sensing, unleashing a wave of innovation that is just beginning to enable creative ways for in situ sensors to draw on nature and their operating environment to ensure that they provide SI-traceable, reliable measurements. The NOAC program is shrinking the precision measurement technology that is currently available only at national metrology institutes like NIST into a suite of sensors that can be embedded directly in equipment or deployed on a factory floor. The following sections describe a few of the nearly two dozen sensors being developed in the program. Chip-Scale Atomic Clocks Precision timekeeping is fundamental to applications in communications, financial transactions, aviation, and GPS. The first chip-scale atomic clock was created by NIST in 2004. It laid the foundation for the NOAC program by showing that it’s possible to shrink a lab-full of complex equipment, which previously required a team of highly trained staff to operate, to a device the size of a grain of rice (NIST 2004)! Chip-scale clocks are now commercially available, and NIST is working on the next generation of them that will operate at optical frequencies (Newman et al. 2019; NIST 2019). These more sensitive clocks have at their heart a vapor cell the size of a coffee bean (figure 1). The clocks could be invaluable in manufacturing synchronization for robot-to-robot communications, monitoring of machine tool life, and overall factory efficiency, among other applications. Photonic Thermometers The current “gold standard” for temperature is the platinum resistance thermometer, which is fragile and vulnerable to humidity, requires expensive calibrations, and is difficult to miniaturize. NIST is developing a replacement (figure 2) based on the relationship between temperature and the optical properties of materials, leading to a sensor that will be inexpensive, small, portable, and robust, with applications to manufacturing process control and instrumentation, among many others. Electric-Field Sensors Advanced manufacturing takes place in electronically noisy environments and it’s critical to characterize both unintended electromagnetic emissions and intentional communication signals sent wirelessly in a dynamic environment. The best commercial instruments are accurate to only within 10 percent. To address this need, NIST has prototyped a fundamentally new type of sensor (figure 3) based on Rydberg atoms, which, because of the highly excited state of their outermost electron, are extremely sensitive to electric fields (Holloway et al. 2014). These prototype sensors have been shown to outperform current sensors in both sensitivity and accuracy (good to about 4 percent), never need to be calibrated, and have the potential to be shrunk to a chip-scale package. Wanted: Fast, Reliable Wireless Connectivity for Harsh Radio-Propagation Environments Automation-driven factories rely on not only a steady stream of accurate sensor information but also uninterrupted and unambiguous connectivity. The growing trend toward augmenting or replacing wired connections with wireless channels comports with the postpandemic need for decentralization and M2M communications. Why Wireless On the factory floor, wireless connectivity offers many benefits over wired solutions, such as the elimination of costly cabling, mobility, configuration flexibility, and improved efficiency in operations (Lee et al. 2020; Liu et al. 2019). Wireless connections can also be used in otherwise impractical locations; for example, low-power monitoring devices can eliminate the difficulties inherent in physically routing cables (Ferreira et al. 2013). The relatively low cost and flexibility of wireless technology infrastructure are especially important for small and midsize factories (Candell et al. 2018). Much of the focus on wireless connectivity for IIoT technology relates to its deployment in workcell-sized environments. A workcell typically consists of a single or a few machines. While the size of factories varies considerably, the size of a workcell is typically on the order of 10 m or less per side (Lee et al. 2020). A focus on requirements for the workcell rather than for the entire factory provides a better picture of user requirements (Candell et al. 2019). Basic assumptions for user requirements in workcells might include (i) no more than one communication system failure every 1000 years; (ii) individual transmissions that are independent, allowing multiple traffic channels to coexist; and (iii) a single wireless link used in certain scenarios involving multiple sensors/regulators (Lee et al. 2020). Recent wireless standards are aimed at providing reliability and latency to meet requirements at the workcell level for the discrete manufacturing sector. For example, IEEE 802.11ax provides simulated latencies on the order of 1–5 ms. However, no wireless standard or technology can yet provide the sub-ms latencies that will be needed for power electronics systems control and other applications that are highly constrained by latency, especially in intraworkcell and intramachine wireless use cases (table 1; Candell and Kashef 2017; Candell et al. 2018). This is because current industrial wireless technologies were, generally, designed for the process manufacturing industry, where sensing and control can successfully happen on the order of seconds. For future discrete manufacturing applications, events between sensors and monitors/controls must be communicated within milliseconds or less. TABLE 1 Current wireless standards (Candell and Kashef 2017; Candell et al. 2018) do not address the needs of workcell communications, which are highlighted in the red (“job-based”), black (“safety”), and blue (“tracking”) boxes. BLOS = beyond line of sight; CSMA = carrier-sense multiple access; LOS = line of sight; RFID = radio frequency identification; TDMA = time-division multiple access; VLBR WAN = very low bit rate wide area networks. Manufacturers have indicated that decentralized wireless would be a flexible and robust means of manufacturing communication for use in intraworkcell and intramachine cases if latencies can be reduced (Candell et al. 2018). For example, for a programmable logic controller (PLC) to provide feedback to an industrial machine based on sensor data, latencies significantly less than the typical PLC scan cycles of 10 ms would be required. How much less would depend on the number of stations and the multiple-access scheme used. Ideally, communication links of 100 sensors would provide ultrareliable factory floor operation. Until resolved technologically, these gaps will affect manufacturers’ ability to trust the viability of decentralized wireless IIoT networks. There is thus great interest in developing robust wireless technology that addresses the unique requirements of sensor communications, latency, reliability, and flexibility in IIoT applications. However, the industrial workcell environment is one of the most challenging for wireless system deployment because of the high reflectivity (including the potential for three-dimensional reflections) and dynamically changing conditions (including intermittent blockage of a channel due to moving elements such as robotic arms or autonomous vehicles). Understanding the range of potential channel impairments is essential for developing robust hardware and network protocols, and is a topic of current research. The coupling of advanced analytics (such as channel estimation) and statistically based channel modeling with accurate real-time assessment will allow the development of wireless hardware that can make use of multiple redundant channels, each with the potential for ultrawideband capacity. Current NIST research uses machine learning to identify and replicate key electromagnetic features of industrial wireless channels for repeatable, laboratory-based testing, which we describe next. Characterizing the Environment in Space and Time As mentioned, one of the foundational building blocks for developing spatial-temporal models of the industrial environment is verified channel-propagation measurements. Such measurements are often carried out in representative environments to study important channel features such as path loss, reflectivity (or multipath), coherence time (how quickly a channel dynamically changes), and spatial characteristics such as the angle of arrival (AoA) of direct and reflected signals. Several research groups are collecting such data in factory environments to facilitate standards development and hardware design. In our research at NIST we use a synthetic aperture–based channel measurement system to characterize the 3D spatial-temporal characteristics of the industrial propagation channel, coupled with machine learning techniques to identify the most typical and prominent features in the environment, as described in the next subsection. Assessing OTA Wireless Device Performance in the Workcell Wireless devices with integrated antennas, such as most IoT devices, require OTA verification of performance. Certification bodies have developed rigorous tests for cell phones and some work has been done on IoT device characterization (CTIA 2020). However, much of this work is focused on whether the device meets radiated power and receiver sensitivity limits in nonreflective, “isotropic” environments, where signals are incident on the device from all angles equally. Newly released standards focus on creating specific channel conditions to test multiple antenna devices, such as cell phones with multiple antennas (3GPP). To support the millisecond timescales needed for adaptive mmWave IIoT, OTA tests must evolve to emulate 3D spatial channel characteristics such as the timing and AoA of reflected signals. As such, NIST is applying a wide range of expertise in wireless communications, manufacturing, and artificial intelligence to create a repeatable, noninvasive dynamic OTA testbed utilizing innovative quantum field probes. The NIST OTA testbed will expose an IIoT device to a repeatable, dynamically changing environment to assess its ability to reconfigure adaptively to changing channel conditions by use of machine learning (Kashef et al. 2021). As an additional benefit for IIoT designers, our measured datasets and models of the dynamically changing channels will be made available for users to design and train their AI-based adaptive network hardware. The concept is illustrated in figure 4. The measured static and dynamic reflective characteristics of the factory floor environment (figure 4a) are extracted, as described above. The channel conditions are then replicated in a lab-based test chamber by placing reflective and electromagnetic wave–absorbing material in the appropriate locations. This process is facilitated by machine learning. Once configured, the chamber characteristics are verified through measurement. The large, metallic bodies of conventional channel sounders perturb the environment in which they operate and are their most important and fundamental accuracy limitation in physically small environments. To address this limitation, NIST is creating minimally reflective quantum field probes to characterize the time-evolving channels in the chamber, leveraging the NOAC Rydberg atom–based probe. An array of the NOAC probes is needed to accurately measure the timing and AoA of signals incident on an IIoT device by sampling the ambient fields at multiple spatial locations in the test chamber through the use of synthetic aperture techniques. The important features of this test setup include the accuracy of the workcell channels that it recreates, including the traceable characterization of these channels with the use of a low-invasiveness probe, and the wide range of channels that can be created. Challenges There are many technical challenges remaining to mature the technologies needed for a reliable, scalable factory infrastructure. Along with those in communications and sensing, outlined above, other challenges to achieving digital transformation include security, culture, and competition. Security IoT and IIoT security is complex, with vulnerabilities in the middle layers of the technology stack, between applications and hardware, across communication channels, and in communication protocols (Friedman and -Goldstein 2019). The security challenge is one of risk management, and the following factors need to be considered and are active areas of NIST research (Lee et al. 2020): identification and authentication control use control system integrity data confidentiality restricted data flow timely response to events resource availability. Culture Embracing new technology—such as self-calibrating quantum-based sensors—requires a cultural shift. Measure-ment assurance is based on an international system of intercomparisons, accreditation, and assessment. Quantum-based sensors can short-circuit this labor-intensive foundation, but only if the global community trusts it. Competition The United States has been slow to realize Industry 4.0. Despite the potential offered by digital transformation, initiatives fail with alarming frequency—a 2016 study by McKinsey showed that 70 percent fail outright (Bucy et al. 2016), and Forbes placed this number as high as 84 percent (Rogers 2016). These false starts may be explained as the tribulations of early adopters. They may also reflect a market failure to institutionalize the infrastructure needed to realize the benefits of digital transformation. While US industry is grappling with its slow climb to success, other countries are investing. According to the Information Technology and Innovation Foundation, other countries are making Industry 4.0 policies a priority by launching “pilot fabs” for smart manufacturing, documenting digitalization use cases (Germany has identified over 300), and providing financial support to industry (Atkinson 2020). Conclusion In this brief overview, we have highlighted ways NIST is working to facilitate a cultural shift in the use of new technology to increase competitive activities in the industrial sector. By addressing some of the measurement-related gaps and supporting standards development for new wireless technology, industry can move forward with greater confidence as new technology becomes available. Through measurements and standards, NIST’s goal is to facilitate the postpandemic adoption of efficient, secure, and decentralized technology for the industrial sector. References Atkinson RD. 2020. The state of US advanced tech industry competitiveness and what to do. Presentation to NIST, Oct 19. Available at https://www.nist.gov/system/files/documents/2020/10/19/ 3.%20Rob%20Atkinson_VCAT_NIST.pdf. Bucy M, Finlayson A, Kelly G, Moye C. 2016. The “how” of transformation. New York: McKinsey & Company. Candell R, Kashef M. 2017. Industrial wireless: Problem space, success considerations, technologies, and future direction. Resilience Week, Sep 18–22, Wilmington DE. Candell R, Hany MT, Lee KB, Liu Y, Quimby JT, Remley CA. 2018. NIST Guide to Industrial Wireless Systems Deployments. Advanced Manufacturing Series 300-4. -Gaithersburg MD: National Institute of Standards and Technology. Candell R, Kashef M, Liu Y, Foufou S. 2019. A SysML -representation of the wireless factory work cell. International Journal of Advanced Manufacturing Technology 104:119–40. CTIA. 2020. Certification: Test Plan for Wireless Large-Form-Factor Device Over-the-Air Performance, v. 2.0. -Washington. Online at https://ctiacertification.org/test-plans/. Ferreira P, Reyes V, Mestre J. 2013. A web-based integration procedure for the development of reconfigurable robotic work-cells. International Journal of Advanced Robotic Systems 10(7):295. Friedman J, Goldstein B. 2019. iNEMI Roadmap, Industrial Internet of Things chapter. Available for purchase from iNEMI (https://www.inemi.org/inemi-2019-roadmap). Holloway CL, Gordon JA, Jefferts S, Schwarzkopf A, -Anderson DA, Miller SA, Thaicharoen N, Raithel G. 2014. Broadband Rydberg atom-based electric-field probe: From self-calibrated measurements to sub-wavelength imaging. IEEE Transactions on Antennas and Propagation 62(12):6169–82. Kashef M, Vouras P, Jones R, Candell R, Remley KA. 2021. Temporal exemplar channels in high-multipath environments. Submitted for IEEE International Conf on -Acoustics, Speech, and Signal Processing, Jun 6–11, Toronto. Lee KB, Candell R, Bernhard H-P, Pang Z, Val I. 2020. Reliable, high-performance wireless systems for factory automation. Internal Report 8317. Gaithersburg MD: National Institute of Standards and Technology. Liu Y, Kashef M, Lee KB, Benmohamed L, Candell R. 2019. Wireless network design for emerging IIoT applications: Reference framework and use cases. Proceedings of the IEEE 107(6):1166–92. Newman ZL, Maurice V, Drake T, Stone JR, Briles TC, -Spencer DT, Fredrick C, Li Q, Westly D, Ilic BR, and 12 others. 2019. Architecture for the photonic integration of an optical atomic clock. Optica 6(5):680–85. NIST [National Institute of Standards and Technology]. 2004. NIST unveils chip-scale atomic clock. NIST News, Aug 27. Gaithersburg MD. NIST. 2019. NIST team demonstrates heart of next--generation chip-scale atomic clock. NIST News, May 17. Gaithersburg MD. Rogers B. 2016. Why 84% of companies fail at digital transformation. Forbes, Jan 7.  https://www.nist.gov/noac  https://www.nist.gov/programs-projects/photonic-thermometry  https://www.ieee802.org/11/Reports/tgax_update.htm  3rd Generation Partnership Project, Technical Specification Group Radio Access Network. Study on channel model for frequencies from 0.5 to 100 GHz (Release 16), 3GPP TR 38.901 V16.1.0 (2019-12). Available at https://www.3gpp.org/. About the Author:Barbara Goldstein is associate director of the Physical Measurement Laboratory and Kate Remley is leader of the Metrology for Wireless Systems Project, both at the National Institute of Standards and Technology.