MFGAN: Multimodal Fusion for Industrial Anomaly Detection Using Attention-Based Autoencoder and Generative Adversarial Network.

Anomaly detection plays a critical role in ensuring safe, smooth, and efficient operation of machinery and equipment in industrial environments. With the wide deployment of multimodal sensors and the rapid development of Internet of Things (IoT), the data generated in modern industrial production has become increasingly diverse and complex. However, traditional methods for anomaly detection based on a single data source cannot fully utilize multimodal data to capture anomalies in industrial systems. To address this challenge, we propose a new model for anomaly detection in industrial environments using multimodal temporal data. This model integrates an attention-based autoencoder (AAE) and a generative adversarial network (GAN) to capture and fuse rich information from different data sources. Specifically, the AAE captures time-series dependencies and relevant features in each modality, and the GAN introduces adversarial regularization to enhance the model's ability to reconstruct normal time-series data. We conduct extensive experiments on real industrial data containing both measurements from a distributed control system (DCS) and acoustic signals, and the results demonstrate the performance superiority of the proposed model over the state-of-the-art TimesNet for anomaly detection, with an improvement of 5.6% in F1 score.

Optimizing Temperature Setting for Decomposition Furnace Based on Attention Mechanism and Neural Networks.

The temperature setting for a decomposition furnace is of great importance for maintaining the normal operation of the furnace and other equipment in a cement plant and ensuring the output of high-quality cement products. Based on the principles of deep convolutional neural networks (CNNs), long short-term memory networks (LSTMs), and attention mechanisms, we propose a CNN-LSTM-A model to optimize the temperature settings for a decomposition furnace. The proposed model combines the features selected by Least Absolute Shrinkage and Selection Operator (Lasso) with others suggested by domain experts as inputs, and uses CNN to mine spatial features, LSTM to extract time series information, and an attention mechanism to optimize weights. We deploy sensors to collect production measurements at a real-life cement factory for experimentation and investigate the impact of hyperparameter changes on the performance of the proposed model. Experimental results show that CNN-LSTM-A achieves a superior performance in terms of prediction accuracy over existing models such as the basic LSTM model, deep-convolution-based LSTM model, and attention-mechanism-based LSTM model. The proposed model has potentials for wide deployment in cement plants to automate and optimize the operation of decomposition furnaces.

CaFANet: Causal-Factors-Aware Attention Networks for Equipment Fault Prediction in the Internet of Things.

Existing fault prediction algorithms based on deep learning have achieved good prediction performance. These algorithms treat all features fairly and assume that the progression of the equipment faults is stationary throughout the entire lifecycle. In fact, each feature has a different contribution to the accuracy of fault prediction, and the progress of equipment faults is non-stationary. More specifically, capturing the time point at which a fault first appears is more important for improving the accuracy of fault prediction. Moreover, the progress of the different faults of equipment varies significantly. Therefore, taking feature differences and time information into consideration, we propose a Causal-Factors-Aware Attention Network, CaFANet, for equipment fault prediction in the Internet of Things. Experimental results and performance analysis confirm the superiority of the proposed algorithm over traditional machine learning methods with prediction accuracy improved by up to 15.3%.

On Fatigue Detection for Air Traffic Controllers Based on Fuzzy Fusion of Multiple Features.

Fatigue detection for air traffic controllers is an important yet challenging problem in aviation safety research. Most of the existing methods for this problem are based on facial features. In this paper, we propose an ensemble learning model that combines both facial features and voice features and design a fatigue detection method through multifeature fusion, referred to as Facial and Voice Stacking (FV-Stacking). Specifically, for facial features, we first use OpenCV and Dlib libraries to extract mouth and eye areas and then employ a combination of M-Convolutional Neural Network (M-CNN) and E-Convolutional Neural Network (E-CNN) to determine the state of mouth and eye closure based on five features, i.e., blinking times, average blinking time, average blinking interval, Percentage of Eyelid Closure over the Pupil over Time (PERCLOS), and Frequency of Open Mouth (FOM). For voice features, we extract the Mel-Frequency Cepstral Coefficients (MFCC) features of speech. Such facial features and voice features are fused through a carefully designed stacking model for fatigue detection. Real-life experiments are conducted on 14 air traffic controllers in Southwest Air Traffic Management Bureau of Civil Aviation of China. The results show that the proposed FV-Stacking method achieves a detection accuracy of 97%, while the best accuracy achieved by a single model is 92% and the best accuracy achieved by the state-of-the-art detection methods is 88%.

On Weather Data-Based Prediction of Gamma Exposure Rates Using Gradient Boosting Learning for Environmental Radiation Monitoring.

Gamma radiation has been classified by the International Agency for Research on Cancer (IARC) as a carcinogenic agent with sufficient evidence in humans. Previous studies show that some weather data are cross-correlated with gamma exposure rates; hence, we hypothesize that the gamma exposure rate could be predicted with certain weather data. In this study, we collected various weather and radiation data from an automatic weather system (AWS) and environmental radiation monitoring system (ERMS) during a specific period and trained and tested two time-series learning algorithms-namely, long short-term memory (LSTM) and light gradient boosting machine (LightGBM)-with two preprocessing methods, namely, standardization and normalization. The experimental results illustrate that standardization is superior to normalization for data preprocessing with smaller deviations, and LightGBM outperforms LSTM in terms of prediction accuracy and running time. The prediction capability of LightGBM makes it possible to determine whether the increase in the gamma exposure rate is caused by a change in the weather or an actual gamma ray for environmental radiation monitoring.

LightFD: Real-Time Fault Diagnosis with Edge Intelligence for Power Transformers.

Power fault monitoring based on acoustic waves has gained a great deal of attention in industry. Existing methods for fault diagnosis typically collect sound signals on site and transmit them to a back-end server for analysis, which may fail to provide a real-time response due to transmission packet loss and latency. However, the limited computing power of edge devices and the existing methods for feature extraction pose a significant challenge to performing diagnosis on the edge. In this paper, we propose a fast Lightweight Fault Diagnosis method for power transformers, referred to as LightFD, which integrates several technical components. Firstly, before feature extraction, we design an asymmetric Hamming-cosine window function to reduce signal spectrum leakage and ensure data integrity. Secondly, we design a multidimensional spatio-temporal feature extraction method to extract acoustic features. Finally, we design a parallel dual-layer, dual-channel lightweight neural network to realize the classification of different fault types on edge devices with limited computing power. Extensive simulation and experimental results show that the diagnostic precision and recall of LightFD reach 94.64% and 95.33%, which represent an improvement of 4% and 1.6% over the traditional SVM method, respectively.

On a Blockchain-Based Security Scheme for Defense against Malicious Nodes in Vehicular Ad-Hoc Networks.

Vehicular ad-hoc networks (VANETs) aim to provide a comfortable driving experience. Sharing messages in VANETs can help with traffic management, congestion mitigation, and driving safety. However, forged or false messages may undermine the efficiency of VANETs. In this paper, we propose a security scheme based on blockchain technology, where two types of blockchain are constructed based on roadside units (RSUs) and Certificate Authorities (CAs), respectively. The proposed security scheme has multifold goals to identify malicious nodes and detect forged messages based on multiple factors, such as reputation of sender nodes, and time and distance effectiveness of messages. In addition, an incentive mechanism is introduced on the RSU blockchain to encourage RSUs to adopt active behaviors. Extensive simulations show that the proposed scheme exhibits superior performances to existing methods in detecting forged messages and identifying malicious nodes. Meanwhile, it provides privacy protection and improves the efficiency of vehicular networks.

On Efficient Deployment of Wireless Sensors for Coverage and Connectivity in Constrained 3D Space.

Sensor networks have been used in a rapidly increasing number of applications in many fields. This work generalizes a sensor deployment problem to place a minimum set of wireless sensors at candidate locations in constrained 3D space to k-cover a given set of target objects. By exhausting the combinations of discreteness/continuousness constraints on either sensor locations or target objects, we formulate four classes of sensor deployment problems in 3D space: deploy sensors at Discrete/Continuous Locations (D/CL) to cover Discrete/Continuous Targets (D/CT). We begin with the design of an approximate algorithm for DLDT and then reduce DLCT, CLDT, and CLCT to DLDT by discretizing continuous sensor locations or target objects into a set of divisions without sacrificing sensing precision. Furthermore, we consider a connected version of each problem where the deployed sensors must form a connected network, and design an approximation algorithm to minimize the number of deployed sensors with connectivity guarantee. For performance comparison, we design and implement an optimal solution and a genetic algorithm (GA)-based approach. Extensive simulation results show that the proposed deployment algorithms consistently outperform the GA-based heuristic and achieve a close-to-optimal performance in small-scale problem instances and a significantly superior overall performance than the theoretical upper bound.

A Fast Algorithm for 2D DOA Estimation Using an Omnidirectional Sensor Array.

The traditional 2D MUSIC algorithm fixes the azimuth or the elevation, and searches for the other without considering the directions of sources. A spectrum peak diffusion effect phenomenon is observed and may be utilized to detect the approximate directions of sources. Accordingly, a fast 2D MUSIC algorithm, which performs azimuth and elevation simultaneous searches (henceforth referred to as AESS) based on only three rounds of search is proposed. Firstly, AESS searches along a circle to detect the approximate source directions. Then, a subsequent search is launched along several straight lines based on these approximate directions. Finally, the 2D Direction of Arrival (DOA) of each source is derived by searching on several small concentric circles. Unlike the 2D MUSIC algorithm, AESS does not fix any azimuth and elevation parameters. Instead, the adjacent point of each search possesses different azimuth and elevation, i.e., azimuth and elevation are simultaneously searched to ensure that the search path is minimized, and hence the total spectral search over the angular field of view is avoided. Simulation results demonstrate the performance characters of the proposed AESS over some existing algorithms.