Special Section on “Advances in Cyber-Manufacturing: Architectures, Challenges, & Future Research Directions”

Cyber-Manufacturing (CM) is a concept inspired by Cyber-Physical Systems (CPS). CM refers to a modern manufacturing system that provides an information-transparent environment to facilitate asset management, provide reconfigurability, and maintain productivity. Recently, many businesses and enterprises involved in manufacturing have realized that digitization is inevitable to remain agile and competitive in the future market. CPS solutions are being widely deployed in CM to solve problems, limit risks, increase efficiency, and obtain factual data to make better decisions. By now, everyone is familiar with this technology, and with 6G driving the future of CPS and CM, it has become vital to prepare for the next generation of applications. The idea of CM stems from the fact that Internet-enabled services have added value to businesses in industries such as retail, music, consumer products, transportation, and healthcare.

Current manufacturing enterprises make decisions using a top-down approach: from overall equipment effectiveness to assigning production requirements without considering the condition of the machines. This typically leads to inconsistencies in operations due to lack of linkage between factories, potential overstocking of spare parts, and unexpected machine breakdowns. Such situations require connectivity between machines as a foundation and, in addition, analytics to transform raw data into information that facilitates user decision making. Expected capabilities of Cyber-Manufacturing Systems (CMS) include machine connectivity and data acquisition, machine health forecasting, fleet-based asset management, and manufacturing reconfigurability.

To learn more about this discipline, in this special issue, we invited researchers and practitioners from academia and industry to present novel and innovative challenges and future research directions in CM. We have focused on some key areas related to CM. These include innovative network architectures, advances in software-defined networks, and edge-enabled architectures. We also share concerns around security and privacy-preserving architectures, as well as key challenges and requirements for future CMS. When it comes to sustainability, efficient resource allocation and energy efficiency are key topics at CM. In addition, this special issue features advances in artificial intelligence and machine learning, as well as disruptive technologies and cloud computing solutions for handling big data. A total of 35 papers were submitted for the special issue, of which 5 were selected for publication—a very strict acceptance rate of 14%. This was done intentionally to ensure that only the strongest contributions that would have a major impact were accepted for this special issue. Next, we present a summary of the papers published in this issue.

In “Tolerance Analysis of Cyber-Manufacturing Systems to Cascading Failures,” Fu et al. look at practical CMS in which node components both relay information and provide services. Due to this dual role, the whole system exhibits the typical physical-services interaction characteristic, which makes CMS more vulnerable to cascading failures than general manufacturing systems. To adequately characterize the cascading process of CMS, the authors of this article first develop an interdependent network model for CMS from a physical-service networking perspective. Based on this, a realistic cascading failure model for CMS is designed, which fully considers the routing-oriented load distribution characteristics of the physical network and the selective load distribution characteristics of the service network. Through extensive experimental analysis, the soundness of the proposed model was verified and some meaningful findings were obtained: (1) attacks on the physical network are more likely to trigger cascading failures and can cause more damage, (2) interdependency failures are the main cause of performance degradation in the service network during cascading failures, and (3) isolation failures are the main cause of performance degradation in the physical network during cascading failures. The obtained results can certainly help users develop a more reliable CMS against cascading failures.

In “Digital Twin of Intelligent Small Surface Defect Detection with Cyber-Manufacturing Systems,” Wu et al. focus on the major technological developments in CPS, where Industry 4.0 has evolved through a significant concept called Digital Twins (DT). However, it is still difficult to establish a relationship between twin simulations and real scenarios taking into account dynamic variations, especially when it comes to the detection of small surface defects with high performance and computational resources. In this work, the authors attempt to construct CMS to provide a DT solution for small surface defect detection. Their proposed system is based on DT and consists of an edge-cloud architecture and a surface defect detection algorithm. The edge-cloud architecture takes into account the dynamic characteristics and real-time response requirements to achieve smart manufacturing by efficiently collecting, processing, analyzing, and storing data produced by factories. A deep learning based algorithm is developed to detect surface defects based on multimodal data (i.e., image and depth data). Experimental evaluation shows that the proposed algorithm achieves high accuracy and recall in detecting small defects, developing DT in CM.

In “Unpaired Self-Supervised Learning for Industrial Cyber-Manufacturing Spectrum Blind Deconvolution,” Deng et al. investigate CM, combining industrial big data with intelligent analysis to find and understand the intangible problems in decision making, which requires a systematic method to handle large-scale signal data. With the development of spectral detection and photoelectric imaging technology, spectral blind deconvolution has achieved remarkable results. However, spectral processing is limited by the one-dimensional signal, and there is no available structural information with a few training examples. Moreover, in most practical applications, it is possible to collect unpaired noise and clean spectrum. Their training method of unpaired learning is practical and valuable. Therefore, in this article, a two-stage deconvolution scheme combining self-supervised learning and feature extraction is proposed to generate two complementary paired sets by self-supervised learning to extract the final deconvolution network. A new deconvolution network is also developed for feature extraction. The spectrum is pre-trained by spectral feature extraction and a noise estimation network to improve the training efficiency and satisfy the assumed noise characteristics. Experimental results show that this method is effective for various types of synthetic noise.

In “Offering Two-Way Privacy for Evolved Purchase Inquiries,” Pennekamp et al. dive into dynamic and flexible business relationships, which are expected to become increasingly important in the future to meet special change requests or small batch production. Today, buyers and sellers must disclose sensitive information about products before they are actually manufactured. Without a relationship of trust, however, this situation is precarious for the companies involved, as they fear for their competitiveness. Related work has so far overlooked this problem: existing approaches only protect the information of a single party, thus hindering dynamic and on-demand business relationships. To address the related research gap of inadequately privacy-protected information and deal with companies without an established trust relationship, the authors pursue innovative privacy-friendly purchasing requests that seamlessly integrate with today's established supplier management and procurement processes. Using established building blocks from private computing, such as private set intersection and homomorphic encryption, the authors propose two designs with slightly different privacy and performance implications to securely implement purchase requests over the Internet. Specifically, the authors allow buyers to consider more potential sellers without disclosing sensitive information, and relieve sellers of the burden of repeatedly generating elaborate but discarded offers. The scalability of the presented approaches is clearly demonstrated using two real-world use cases from the manufacturing engineering domain. Overall, deployable concepts are presented that enable two-way privacy for purchase inquiries and, in turn, fill a gap that currently hinders establishing dynamic and flexible business relationships. In the future, research activities in this overlooked area are expected to increase significantly to meet the demands of an evolving manufacturing landscape.

Finally, for an invited paper of “Exploring the Potential of Cyber Manufacturing Systems in the Digital Age,” Ahmed et al. focus on Cyber Manufacturing Systems (CMS), which are becoming increasingly popular. CMS are creating a move from conventional manufacturing to an innovative paradigm that emphasizes innovation, automation, better customer service, and intelligent systems. A new manufacturing model can improve efficiency and productivity and enable better customer service and faster response times. It can also revolutionize the way products are made, from design to completion. Therefore, it is likely that this new manufacturing model will become mainstream in the near future. By building new technologies on top of existing CMS, these systems ensure that data exchange and integration between decentralized systems are reliable and secure. Recently published case studies from industry and the literature support this claim. There are still some challenges to overcome, such as ensuring data reliability, but these can be overcome with further research and development. In summary, the use of CMS can revolutionize the manufacturing industry. This article provides a comprehensive analysis of these systems and their potential applications and impacts. The article also provides an overview of the field and then goes into more detail about the various aspects of CMS. A taxonomy of the most common and current approaches to CMS is presented, including networked CMS, distributed CMS, cloud-based CMS, and CPS. In addition, the authors identify several popular open source software and datasets, and discuss how these resources can reduce barriers to CMS research. The article also identifies several important issues and research opportunities related to CMS, including better integration of hardware and software, improved security and privacy protocols, communication protocols, and improved data management systems. The authors provide a comprehensive overview of current technology and valuable insights into the potential impact of the technology on society and industry.

The articles published in this special issue will be instructive and inspiring for our readers. Last but not least, the guest editors would like to thank Arriane Bustilo for spending countless hours on this special issue, answering questions and providing feedback to us as editors when needed. Finally, we would like to thank Professor Ling Liu, Editor-in-Chief of ACM Transactions on Internet Technology (TOIT), for this invaluable opportunity and collaboration in finalizing this special issue, as well as all authors and reviewers for their contributions to it.