Reconstruction-based anomaly detection is appealing for industrial inspection because it reconstructs anomaly-free references and produces interpretable pixel-level residual maps. However, in multiclass settings, it often suffers from, first, identity mapping, where abnormal regions are overreconstructed and residuals vanish, and second, cross-category feature entanglement, which introduces artifacts and weakens localization. Residual scoring is further biased because it mixes true defect discrepancies with model-induced reconstruction errors, causing false alarms and confusing explanations. We propose a semantically guided counterfactual model (SGCAD) for multiclass anomaly detection and localization. SGCAD applies a masking intervention to a candidate region, reconstructs the intervened input, and contrasts factual and intervened residuals to estimate reconstruction bias, yielding a debiased anomaly map by treating consistent differences as anomalies. To stabilize reconstruction, we introduce a dual-branch bottleneck with latent constraints to mitigate identity mapping while preserving normal details, and a hierarchical adaptive prompt to reduce cross-category confusion during decoding without class priors. Extensive experiments on four public benchmarks demonstrate that while SGCAD maintains competitive image-level detection performance, it achieves superior state-of-the-art results in pixel-level localization. By effectively mitigating bias-induced false positives, SGCAD offers a more precise and reliable solution for practical industrial inspection.

In this article, in the occurrence of load disturbances and internal actuator faults, exact performance compensation problem for nonlinear interconnected power systems (NIPSs) integrated with wind farms is solved by presenting an observer-based sliding mode control strategy. Concretely, a dynamic event condition in the sensor channel is designed by virtue of triggered output error and a dynamic threshold. With the help of event outputs and neural networks (NNs) technology, a composite observer is proposed to perform state, disturbance, and fault estimations of the power system simultaneously. By utilizing the variable estimations, a novel sliding mode manifold is designed and an observer-based controller is derived directly. Intuitively, the compensation action is triggered once the unknown variables of the NIPSs are estimated, thereby the disturbance rejection and fault tolerance can be achieved. A prominent benefit of the presented strategy is that the estimation convergence ability of the observer is proved independently, then the relationship between the observer and controller for the NIPSs can be relaxed. The control ability is analyzed within aH∞performance through the Lyapunov theory. Simulation results for the two area NIPSs and comparisons on the recent achievements of the NIPSs are made, the results show that more than ten percent improvements can be realized compared with the recent control methods, which in return demonstrate the compensation performance of the presented control strategy.

In low-light environments, images often suffer from quality degradation issues such as low contrast, insufficient brightness, and color distortion due to inadequate light reaching the camera sensor. Most existing methods overlook the issue of color distortion caused by insufficient illumination when enhancing low-light images. In this work, we propose GWRetinex-Net, a novel deep learning network model based on Retinex theory and the gray world assumption. It consists of an image decomposition network, a reflection component enhancement network, and an illumination component enhancement network. During the training process of the image decomposition network, the gray world assumption is innovatively introduced to constrain the illposed decomposition problem caused by the absence of ground truth for reflection and illumination components, ensuring that the reflection components obtained after decomposition retain accurate color information. Based on the decomposition results, the reflection component enhancement network is responsible for mitigating degradation in the reflection components of low-light images, while the illumination component enhancement network focuses on adjusting the illumination distribution in the illumination components of low-light images. Comprehensive qualitative and quantitative experiments demonstrate that our GWRetinex-Net significantly outperforms comparative methods on multiple public datasets. Compared with the best-performing comparative methods, the images enhanced by the proposed method achieve an average improvement of 6.55% in SSIM and 8.87% in PSNR, along with an average decrease of 14.68% in MAE and 6.10% in NIQE. Additionally, object detection experiment in low-light environments further reveals the potential application value of GWRetinex-Net.

In this article, we focus on the issue of adaptive fuzzy finite-time tracking control for a class of uncertain fractional-order nonlinear systems with external disturbance. A new finite-time fractional-order command filtered implementation scheme is presented for the adaptive backstepping method. In the command filtered implementation approach, analytical computation of the fractional derivatives of the stabilizing functions is not necessary. Therefore, the controller and adaptive law of the systems are easier to derive and implement. Based on the proposed command filter, adaptive backstepping technique, and fractional Lyapunov’s direct method, a novel adaptive fuzzy finite-time controller is designed, which can guarantee that the tracking error converges to a small neighborhood of the origin in finite time, and other signals of the closed-loop systems are semiglobal uniform ultimate bounded. In the design process of the controller, fuzzy logic systems (FLSs) are employed to approximate the unknown nonlinear functions, and an auxiliary function is adopted to compensate for the unknown external disturbance and approximation errors of FLSs. Finally, a simulation example is given to show the effectiveness and availability of the proposed control strategy.

This article proposes a new multiobjective optimization approach for designing a self-generated interpretable fuzzy logic system (FLS). The types of fuzzy sets (FSs) can be constructed automatically by self-organizing method, so as to form a hybrid fuzzy system. Different from the existing evolutionary type-1 fuzzy system, which is full of type-1 FSs, and the evolutionary interval type-2 fuzzy system, which is full of interval type-2 FSs, there are both type-1 FSs and interval type-2 FSs in the hybrid fuzzy system. A new transparency-oriented objective function is defined, and the constraint of the footprint of uncertainty of the interval type-2 (IT2) FS is considered for the first time. A new FS merging criterion focusing on the proximity of the cores of FSs is proposed, which is easy to calculate and maintains the characteristics of classical similarity measures. Combined with the new merging criterion, the online cluster and FS updating algorithm is employed to initialize the reference rule base and the type of FS, as it is assumed that no training data are collected in advance. Based on the reference rule base, the advanced multiobjective front-guided continuous ant colony optimization algorithm is introduced to optimize all the free parameters of the FLS. With the operation mentioned above, the self-generated FLSs achieve a good balance between interpretability and performance. The effectiveness of the proposed method is verified by three nonlinear system tracking problems.

In this article, a novel design of a hierarchicalfuzzy system (HFS) based on a self-organized fuzzy partition and fuzzy autoencoder is proposed. The initial rule set of the system is empty, and all the fuzzy sets and fuzzy rules are generated by a self-organized fuzzy partition algorithm. By adopting an improved box plot data standardization method, the processed data can more accurately represent the distribution characteristics of the input data, which improve the accuracy and the rationality. A fuzzy autoencoder is used to train the HFS layer by layer, which can not only ensure the effectiveness of the fuzzy system's hidden layer variables but also provide interpretability. Compared with the traditional fuzzy logic system, the HFS reduces the total number of rules and the complexity. The proposed HFS is tested on three different regression datasets. The experimental results illustrate that the hierarchical self-organized fuzzy system still performs better in terms of regression accuracy indicators than the self-organized fuzzy system.

Robust stability and stabilization problems for a class of discrete-time nonlinear systems are studied in this article. The control signal is transmitted to the nonlinear systems through a lossy communication channel in which time-varying network-induced delay occurs. A Takagi-Sugeno (T-S) fuzzy model is used to represent the nonlinear systems. Using membership function boundary information, a novel membership-function-dependent (MFD) analysis approach is presented. Unlike most previous results, the proposed MFD analysis approach does not need to increase the number of slack variables to reduce conservatism. Combining auxiliary matrices and this approach, we propose some new fuzzy Lyapunov-Krasovskii functionals (FLKFs) that do not require the restrictions of positive definiteness imposed on Lyapunov matrices. Using the FLKF and MFD analysis approach, a sufficient condition in the form of linear matrix inequalities (LMIs) is developed to guarantee the stability of T-S fuzzy time-varying delay systems. Based on the results for stability, less conservative stabilization conditions are also derived. Finally, numerical examples show that less conservative results can be obtained by the proposed approach. The inverted pendulum system with state delay or network-induced delay is also studied to verify the effectiveness of the suggested approach.

In this article, a mathematical model of a magnetic-wheeled mobile robot (MWMR) used for wall climbing in some special industrial sites is established, and to realize the precise motion control of the robot, an intelligent discrete algorithm for trajectory tracking control of the MWMR is presented. The robot is subjected to nonholonomic constraints when moving on the wall. The discrete mathematical model of the MWMR is established with the description of the kinematics and dynamics, where the dynamics are described by the second-order Lagrange's equations. Improved dual-heuristic dynamic programming (DHP) where the actor-critic structure adopts random vector functional link neural networks (RVFL NNs) is the main configuration of the trajectory tracking control algorithm. Moreover, the tracking control algorithm is supplied by a PD controller and a supervisory element to generate overall control signals. The improved strategy is to optimize the input layer weights of RVFL NNs by a genetic algorithm to improve the approximation performance of DHP. Simulations are performed to test the trajectory tracking control algorithm for this wall-climbing robot, and the results are compared with those of neural network tracking control algorithm. Comparative analysis verifies the effectiveness and advancement of the proposed method.

This study explores the stabilization problem of nonlinear systems subject to parameter uncertainties, time-varying delay, and data-packet dropouts. An interval type-2 Takagi-Sugenofuzzy model is employed to represent nonlinear systems subject to parameter uncertainties. The lower and upper membership functions of the system enable parameter uncertainties to be captured effectively. To enhance the flexibility of design, the proposed state feedback controller does not share the premise membership functions of the model. Unlike the majority of existing results, the present study simultaneously considers parameter uncertainties, time-varying delay, and data-packet dropouts in nonlinear control systems. A novel finite-sum inequality that includes well-known inequalities as special cases, is developed to derive new stability conditions in the form of linear matrix inequalities. Several examples of simulation are used to demonstrate the effectiveness of the proposed approach.