Two-photon lithography (TPL) is an advanced technique used for additive manufacturing. How to effectively inspect the part quality is one of the challenges of TPL before large-scale industrial application. To produce cured part, the light dosage parameter is limited during the fabrication process, and the limit varies from different application scenarios. By automatic recognition of part quality, engineers can efficiently find light dosage limits and monitor the fabrication process. This paper introduces a visual monitoring-based video Transformer with three-dimensional (3D) shifted window multi-head self-attention for automatically detecting part quality in four typical real scenarios. This framework introduces a multi-head self-attention mechanism to capture global features, thereby integrating spatial and sequential information for part quality recognition. The 3D shifted window mechanism is also applied to introduce the locality similar to convolution and reduce computational complexity. In addition, hierarchical representation is introduced to Transformer architecture, which helps to model high-level information from low-level features. The dataset with four scenarios, which are different in write pattern and photoresist, is used to evaluate the feasibility of the industrialization of this framework. The results show that the proposed method has better performance than the traditional deep learning model in the detection of part quality.

In the long-term operation process, rolling bearings often have multiple single faults or cascade faults, which are coupled with each other to form the compound fault. The complex coupling components of compound fault lead to difficultly building a correlation between compound fault feature and fault class. Moreover, for the transfer learning based on domain adaptation, compound fault feature of source domain is pretty different from that of target domain, thus knowledge of source domain is hard to be transferred into the cross-domain unsupervised learning process of target domain. To solve above mentioned problems well, this paper proposes a domain reinforcement feature adaptation methodology with correlation alignment (CA-DRFA) to complete the cross-domain compound fault diagnosis of bearings. Specifically, a deep reinforcement learning model is improved by being combined with the domain adversarial training way, and meanwhile the correlation alignment metrics is adopted to enhance the generalization performance in feature alignment. In addition, two experiments and an engineering application are performed to demonstrate that CA-DRFA outperforms other popular cross-domain fault diagnosis methods in cross cutting condition and speed transfer tasks. Overall, CA-DRFA is able to implement cross-domain compound fault diagnosis with high accuracy, and effectively reduce the cost of labeling target samples.

Natural gas pipeline system operations rely heavily on Supervisory Control and Data Acquisition (SCADA) systems. While the SCADA systems introduce many advantages, they also introduce more vulnerabilities by providing opportunities for malicious cyber-attackers. If the cyber-attacks properly modify pressures, flows, and the topology the operator believes is present simultaneously, the cyber-attacks can be undetectable. While this topic has received attention for electrical grids, other cyber-physical systems have seen much less study on this topic. Natural gas networks are employed extensively to power generators in the electrical grid, so attacks on natural gas networks are very important. We have not seen any research on this topic for natural gas networks yet. The particular nonlinear equations which model natural gas networks make the analysis much more difficult. In this paper, we study undetectable attacks on natural gas networks in a signal processing perspective by describing the steady-state mathematical model and sensor measurements. We propose a countermeasure to eliminate undetectable attacks by protecting sensors in specific locations. We present an example that describes how an operator can be misled if the proposed countermeasure is not applied. In such cases, the operator could apply inappropriate control which could damage the system or cause a loss of critical gas supply to customers.

Sensor-equipped connected vehicle networks (SECVNs) have the potential to enable substantially safer driving by improved object tracking, which is an important basic building block in SECVNs. Unfortunately, cyber-attacks on SECVNs pose a very serious threat which could lead to unacceptable outcomes, including fatalities. Recently there has been increasing focus on malicious attack detection and mitigation in SECVNs, and some of this work has considered attacks on sensor data to impact object tracking. Unfortunately, low complexity mitigation approaches which do not compromise performance are lacking. This paper describes an efficient machine-learning enhanced approach for tracking under cyber-attacks. By proper selection of some variances related to the sensor and prior probability density functions, under some assumptions the performance can be made as close as desired to a bound on the best possible performance. However, the complexity of this new approach is dramatically lower than the best existing published low complexity approach, which provides performance which is substantially inferior to that provided by the new approach. The new approach also provides much better scaling with the size of the SECVN. In particular, the complexity increases linearly in the number of sensors, while the best low complexity published approach has a complexity which grows quadratically in the number of sensors. The new approach is also applicable to other tracking applications.

Accurate knowledge of natural gas network topology is critical for the proper operation of natural gas networks. Failures, physical attacks, and cyber attacks can cause the actual natural gas network topology to differ from what the operator believes to be present. Incorrect topology information misleads the operator to apply inappropriate control causing damage and lack of gas supply. Several methods for verifying the topology have been suggested in the literature for electrical power distribution networks, but we are not aware of any publications for natural gas networks. In this paper, we develop a useful topology verification algorithm for natural gas networks based on modifying a general known statistics-based approach to eliminate serious limitations for this application while maintaining good performance. We prove that the new algorithm is equivalent to the original statistics-based approach for a sufficiently large number of sensor observations. We provide new closed-form expressions for the asymptotic performance that are shown to be accurate for the typical number of sensor observations required to achieve reliable performance.