Data-Centric Metrology in Future Manufacturing

Gregory W. Vogl¹, Aaron W. Cornelius¹ and Xiaodong Jia²

1 Engineering Laboratory, National Institute of Standards and Technology, Gaithersburg, USA 2 Department of Mechanical and Materials Engineering, University of Cincinnati, Cincinnati, USA

Status

In modern smart factories, process data is collected and logged throughout the entire manufacturing process [1]. This creates huge datasets which include machine settings and process parameters, equipment sensor data, metrology results, and maintenance logs. Data-centric metrology (DCM) aims to leverage this collected data to provide more accurate estimates for part quality, reduce the overall measurement cost, and provide information necessary to optimize manufacturing [2]. While external information has long been used to improve measurement quality at a basic level (e.g., compensating measurements for temperature changes), DCM will actively monitor, optimize, and adapt measurements as required. With advancements in robotics and automation, DCM is poised to become a key enabler of future intelligent metrology technologies with (semi-) automated decision-making abilities to enhance precision and efficiency. DCM combines three main pillars which have seen independent research:

(1) Integrated metrology is the incorporation and exploitation of metrologically traceable data within manufacturing systems [3]. These on-machine measurements include not just part features but also machine performance, providing real-time feedback on part quality and process health [4].

(2) Virtual metrology is used to estimate part quality for features or process parameters which cannot be directly measured in situ. This is done using digital twins and models which incorporate what process and metrology data is available to estimate missing parameters [5]. (3) A data management system collects all available metrology information to track part quality, provide real-time uncertainty estimates for measurements and system behaviour, and suggest actions to improve measurement performance, e.g., scheduling additional measurement cycles when virtual metrology uncertainty is too high or flagging unreliable sensors for maintenance [6]. These three areas are unified using models based on artificial intelligence (AI) to help parse and act upon the vast volumes of collected data. The semiconductor industry offers perhaps the best look into the future potential for DCM. Process data is used to generate real-time defect estimates and select wafers for further inspection [6]. The process data is then collected and analysed by machine learning (ML) tools to help operators understand and improve processes. However, the transition to smart manufacturing is not uniform: many industries lag behind and are not well-positioned to implement DCM [7]. There are significant technical gaps that must be filled to make widespread deployment of DCM practical and trusted.

Current and future challenges

Metrology is critical for production as the key to process and quality control, and new developments must be thoroughly validated to raise manufacturer confidence in data-driven metrology and drive adoption. The following challenges currently restrict the viability of DCM:

(1) Integrated metrology Integrated metrology and in situ measurements are challenging to perform. In process measurements cannot interfere with the manufacturing process, but at the same time the measurements may be affected by the process since they occur in the same workzone [1] and may encounter various uncontrollable variations [8]. The measurements must also keep pace with the manufacturing process, further restricting what measurements are feasible to perform in situ. As a result, in one survey only 38% of companies performed in-situ measurements [1]. New developments are necessary to create sensors which can deliver low-uncertainty results, at an acceptable pace, and in a variety of environmental conditions. (2) Virtual metrology It is critical for human operators to provide their expertise and maintain visibility of the system health, which is difficult as the number of sensor data streams and automated decisions increases. Hence, new methods must be developed to help users rapidly digest, evaluate, and act upon large amounts of process data [9]. One likely path is the use of AI for automation, virtual metrology, and dynamic sampling of metrological data. Challenges for practical virtual metrology include the creation of an effective initial model using historical data and self-learning updating of models using online data [10]. (3) Data management with uncertainty quantification The future of manufacturing depends upon secure, searchable, scalable, and standardized data architectures in which digitized information from all levels of production will enable real-time adjustements, e.g., with language-neutral identifiers and standardized machine-readable SI formats [11]. Also, data systems should be secure against cyberattacks since increasing connectivity has contributed to dramatic increases in the number of cyberattacks [12]. Challenges towards applying AI-driven insights across the product lifecycle include the curation of big data, interpretation and trust of AI-driven results [13], automatic updating of AI-based models, privacy-preserving methods, and robustness to both class imbalances [14] and variable data quality [15]. AI technologies are often difficult to generalize for deployment, since most AI/ML methods

require significant training data and still may not work as intended in a different setting [13]. To gain user confidence, it is therefore imperitive to provide quantifiable uncertainty estimates for AI-based models [13].

Advances in science and technology to meet challenges

Figure 1 shows a roadmap to achieve data-centric metrology based on the three main pillars of DCM;

(1) integrated metrology, (2) virtual metrology, and (3) data management with uncertainty quantification: (1) Integrated metrology Since traceability is difficult for integrated sensors that cannot be easily removed, new methods should be developed for in situ verification and calibration with traceability to international standards. To facilitate “hot-swapping” of poorly performing sensors, instruments can communicate real-time performance estimates to a centralized measurement management system to trigger verification cycles and flag sensors for repair or replacement. Smart sensors, which are sensors with custom ML inferences, may also be integrated into chips for real-time measurements of chip health [16]. Methods of traceability and calibration may be incorporated via calibration artifacts, self-calibration methods, and standardized processing of metrology-specific data [17], e.g., for robot-assisted metrology with fully automatic data handing. (2) Virtual metrology Data sampling rates within digital twins should be based on AI-driven intelligence to measure the “right amount” of data and minimize the cost of data collection and storage while maintaining product quality. For example, whenever a real-time, AI-estimated uncertainty exceeds a threshold, a measurement may be triggered to gain a data point and minimize the uncertainty at that moment while adding additional data for updating the model. DCM leverages the pattern-learning nature of AI with the trustworthiness of metrology to create trusted, yet machine-unique, models for process control [18]. Periodic comparisons of real-time traceable measurements and model estimations will help quantify the uncertainty of AI-based models. Also, the challenge of an initial model may be aided by transfer learning [10] with an initially heavy dependence on integrated metrology that lessens as the machine-specific model is learned over time. (3) Data management with uncertainty quantification A ubiquitous standardized data architecture is needed for all manufacturing data which validates data quality and provenance, e.g., based on OPC-UA and the digitalization of calibration reports via digital calibration certificates [19]. Methods for quantifying the total output uncertainty of AI-based algorithms, including the inherent uncertainties of the learned model and the input data uncertainties, should be developed and internationally standardized, similar to the GUM [20]. Uncertainties should be estimated to enable dynamic sampling [6] and the propogation of uncertainties, such as with a Shapley Additive exPlanations (SHAP)-based human-readable explainable AI framework [15]. Fully automated data stream handling with low computational latency presents another major challenge for DCM, requiring the innovations in Internet of Things (IoT) hardware and hardware-software optimization.

Figure 1. Roadmap for data-centric metrology (DCM) in future manufacturing based on the three main pillars of DCM: (1) integrated metrology, (2) virtual metrology, and (3) data management with uncertainty quantification.

Concluding remarks

Data-centric metrology can improve manufacturing via reduced measurement costs and increased information for process optimization. Integrated metrology is used to take measurements on-machine during the manufacturing process, virtual metrology uses AI to estimate part quality based on in-process information that cannot be measured, and a data management platform uses all logged data and information to track part quality and provide uncertainty estimates. The future of DCM in manufacturing will address all challenges, e.g., via a standardized data architecture and sampling rates based on uncertainties. Ultimately, data-centric metrology will enable AI to become a trusted extension of human intelligence in manufacturing.

References

[1] Leach RK 2020 Integrated Metrology 10-Year Roadmap for Advanced Manufacturing. High Value Manufacturing Catapult: 1-

38. https://hvm.catapult.org.uk/wp-content/uploads/2022/06/Integrated-Metrology-10-Year-Roadmap.pdf [2] Lee J, Kundu P, Gore P 2022 Industrial AI-Augmented Data-Centric Metrology for Highly Connected Production Systems. Smart and Networked Manufacturing: Wiener Produktionstechnik Kongress 2022. Vienna, Austria: new academic press. 20-3. https://www.newacademicpress.at/gesamtverzeichnis/fertigungstechnik/smart-and-networked-manufacturing-ebook/ [3] Archenti A, Gao W, Donmez A, Savio E, Irino N 2024 Integrated Metrology for Advanced Manufacturing. CIRP Annals 73 639-

39. https://doi.org/10.1016/j.cirp.2024.05.003 [4] Gunasegaram DR, Barnard AS, Matthews MJ, Jared BH, Andreaco AM, Bartsch K, et al. 2024 Machine Learning-Assisted in-Situ Adaptive Strategies for the Control of Defects and Anomalies in Metal Additive Manufacturing. Additive Manufacturing 81 104013. https://doi.org/10.1016/j.addma.2024.104013 [5] Zhang Y, Li L, Yu Q 2024 Virtual Metrology for Enabling Zero-Defect Manufacturing: A Review and Prospects. The International Journal of Advanced Manufacturing Technology 130 3211-27. https://doi.org/10.1007/s00170-023-12726-x [6] Han X, Miller M, Moyne J, Vogl GW, Penkova A, Jia X 2025 A Comparative Study of Semiconductor Virtual Metrology Methods and Novel Algorithmic Framework for Dynamic Sampling. IEEE Transactions on Semiconductor Manufacturing 38 232-9. https://doi.org/10.1109/TSM.2025.3531920 [7] World Economic Forum 2022 The Global Smart Industry Readiness Index Initiative: Manufacturing Transformation Insights Report 2022. World Economic Forum: 1-38. https://www3.weforum.org/docs/WEF_The_Global_Smart_Industry_Readiness_Index_Initiative_2022.pdf

[8] IEEE 2023 International Roadmap for Devices and Systems™: 2023 Update-Metrology. Institute of Electrical and Electronics Engineers (IEEE): 1-67. https://irds.ieee.org/images/files/pdf/2023/2023IRDS_MET.pdf [9] Filz M-A, Bosse JP, Herrmann C 2024 Digitalization Platform for Data-Driven Quality Management in Multi-Stage Manufacturing Systems. Journal of Intelligent Manufacturing 35 2699-718. https://doi.org/10.1007/s10845-023-02162-9 [10] Hsieh YM, Wang TJ, Lin CY, Tsai YF, Cheng FT 2022 Convolutional Autoencoder and Transfer Learning for Automatic Virtual Metrology. IEEE Robotics and Automation Letters 7 8423-30. https://doi.org/10.1109/LRA.2022.3187617 [11] Barbosa CRH, Sousa MC, Almeida MFL, Calili RF 2022 Smart Manufacturing and Digitalization of Metrology: A Systematic Literature Review and a Research Agenda. Sensors 22 6114. https://doi.org/10.3390/s22166114 [12] Mekala SH, Baig Z, Anwar A, Zeadally S 2023 Cybersecurity for Industrial IoT (IIoT): Threats, Countermeasures, Challenges and Future Directions. Computer Communications 208 294-320. https://dx.doi.org/10.1016/j.comcom.2023.06.020 [13] Case Western Reserve University 2024 A Manufacturing Technology Roadmap for AI-Enhanced Multimodal Sensing of Materials and Processes for Complete Product Lifecycle Performance. Case Western Reserve University: 1-60. https://case.edu/issacs/sites/default/files/2024- 02/A_Manufacturing_Technology_Roadmap_for_AI_Enhanced_Multimodal_Sensing_of_Materials_and_Processes_for_Complete_ Product_Lifecycle%20_Performance.pdf [14] Cerquitelli T, Pagliari DJ, Calimera A, Bottaccioli L, Patti E, Acquaviva A, et al. 2021 Manufacturing as a Data-Driven Practice: Methodologies, Technologies, and Tools. Proceedings of the IEEE 109 399-422. https://doi.org/10.1109/JPROC.2021.3056006 [15] Blair J, Amin O, Brown BD, McArthur S, Forbes A, Stephen B 2024 The Transfer Learning of Uncertainty Quantification for Industrial Plant Fault Diagnosis System Design. Data-Centric Engineering 5 e41. https://doi.org/10.1017/dce.2024.54 [16] David J, Gupta A, Bamb R, Akar S, Holt J, Strojwas A, et al. 2024 New Paradigm of Yield Analysis in Big Data and AI Era in Semiconductor Manufacturing. 2024 8th IEEE Electron Devices Technology & Manufacturing Conference (EDTM). https://doi.org/10.1109/EDTM58488.2024.10512106 [17] Catalucci S, Thompson A, Piano S, Branson DT, Leach R 2022 Optical Metrology for Digital Manufacturing: A Review. The International Journal of Advanced Manufacturing Technology 120 4271-90. https://doi.org/10.1007/s00170-022-09084-5 [18] Pandhare V, Miller M, Vogl GW, Lee J 2023 Ball Screw Health Monitoring with Inertial Sensors. IEEE Transactions on Industrial Informatics 19 7323-34. https://doi.org/10.1109/TII.2022.3210999 [19] Grasso Toro F, Lehmann H 2021 Brief Overview of the Future of Metrology. Measurement: Sensors 18 100306. https://doi.org/10.1016/j.measen.2021.100306 [20] Eichstädt S, Keidel A, Tesch J 2021 Metrology for the Digital Age. Measurement: Sensors 18 100232. https://doi.org/10.1016/j.measen.2021.100232