In this study, a simple decentralized H$ _{\infty} $ fault-tolerant observer-based proportional-integral-derivative formation tracking design is proposed for network control systems of large-scale low earth orbit satellites under external disturbance, coupling and malicious attack signals via wireless communication channels. First, a novel reference-based feedforward linearization control scheme is introduced, transforming the nonlinear formation output feedback tracking control problem into an equivalent linearized formation tracking control system of each satellite. To prevent faults from corrupting the estimation and control of the satellite formation, two novel smoothing models of actuator and sensor fault signals are embedded in the equivalent linearized formation system of each satellite. Then, a decentralized H$ _{\infty} $ fault-tolerant observer-based proportional-integral-derivative control strategy is proposed to efficiently attenuate the effect of actuator and sensor faults, measurement noise and satellite coupling on the overall team formation. We only need to solve a linear matrix inequality-constrained optimization problem for each satellite to achieve the optimal H$ _{\infty} $ formation problem. Finally, a team formation example with twelve satellites crossing four orbits for a specific mission is provided to validate the proposed design, comparing it with other methods.

The rapid advancements in deep learning have significantly transformed the landscape of autonomous driving, with profound technological, strategic, and business implications. Autonomous driving systems, which rely on deep learning to enhance real-time perception, decision-making, and control, are poised to revolutionize transportation by improving safety, efficiency, and mobility. Despite this progress, numerous challenges remain, such as real-time data processing, decision-making under uncertainty, and navigating complex environments. This comprehensive review explores the state-of-the-art deep learning methodologies, including Convolutional Neural Networks (CNNs), Recurrent Neural Networks, Long Short-Term Memory networks, and transformers that are central to autonomous driving tasks such as object detection, scene understanding, and path planning. Additionally, the review examines strategic implementations, focusing on the integration of deep learning into the automotive sector, the scalability of artificial intelligence-driven systems, and their alignment with regulatory and safety standards. Furthermore, the study highlights the business implications of deep learning adoption, including its influence on operational efficiency, competitive dynamics, and workforce requirements. The literature also identifies gaps, particularly in achieving full autonomy (Level 5), improving sensor fusion, and addressing the long-term costs and regulatory challenges. By addressing these issues, deep learning has the potential to redefine the future of mobility, enabling safer, more efficient, and fully autonomous driving systems. This review aims to provide insights for stakeholders, including automotive manufacturers, artificial intelligence developers, and policymakers, to navigate the complexities of integrating deep learning into autonomous driving.

This paper investigates the adaptive neural control of vehicular platoons subject to unknown nonlinear functions and full-state constraints. To address the challenges posed by unknown functions, the neural network technology is integrated into the backstepping control framework. Additionally, the time-varying constraints on position, velocity, and acceleration are effectively managed through the application of tangent barrier Lyapunov functions. Notably, the proposed approach successfully avoids the singularity problem. Based on Lyapunov stability theory, it is rigorously shown that the closed-loop system remains bounded, with system states and error signals strictly confined within the prescribed constraint boundaries. Finally, a numerical example is presented to validate the effectiveness and feasibility of the proposed control scheme.

To address the issues of detail loss and blurred restoration in the non-uniform image dehazing process of existing non-uniformly hazy images, this paper presents a novel non-uniform image dehazing algorithm based on serialized integrated attention and multi-dimensional Transformer. This approach aims to restore clear, detailed scenes from heavily hazy images. Firstly, a serialized integrated attention module is established to capture image features. This module amalgamates spatial and channel attention mechanisms and is applied to the shallow-layer network. It effectively concentrates on the local features of the image in both spatial and channel dimensions. Secondly, a multi-dimensional Transformer module is incorporated into the deep-layer network to extract global information and reduce information loss during feature extraction. Finally, feature network fusion is carried out to adaptively fuse the feature maps of the shallow layer and the deep layer. This allows the model to take into account local and global information, combine the detailed local features of the shallow layer with the broad global information of the deep layer, and capture fine-grained details while integrating the image context. The experimental results clearly demonstrate the effectiveness of the proposed algorithm. On the Ⅰ-HAZE, O-HAZE, and NH-HAZE non-uniform haze datasets, the algorithm achieves Peak Signal-to-Noise Ratio values of 22.86, 25.86, and 22.06, along with Structural Similarity Index Measurement values of 0.8731, 0.7799, and 0.7796, respectively. Moreover, the effectiveness of this algorithm is verified on real-world hazy images. Compared with other dehazing algorithms, our proposed method outperforms them in both visual effects and objective metrics.

In this paper, a dual-channel event-triggered control (ETC) protocol with a state estimator is designed in multi-agent systems with switching topologies under denial-of-service attacks. Firstly, an ETC protocol and a dynamic ETC protocol are designed in the communication channel and the controller–actuator channel, respectively. Different from the traditional single-channel ETC, the dual-channel ETC is designed to further save resources. Second, an estimator is introduced to avoid continuous communication between agents. The sufficient conditions for realizing consensus are obtained under denial-of-service attacks. Moreover, due to the unstable communication topology of multi-agent systems, we designed a distributed controller based on switching topologies. Finally, the feasibility of the proposed method is verified through numerical simulations.