Maglev vehicles, which can be regarded as large vehicle models, have tens of thousands of electronic components. The security problem of large systems has always been a challenge. If partial actuator failures occur in these components, it will lead to critical safety problems for vehicles traveling at high speed. In this paper, a fault-tolerant control strategy is developed to solve the problem of partial actuator failures in the levitation module control system of maglev vehicles based on multi-models. First, a dynamic model for the multiple-point levitation module of a maglev vehicle is established. Considering the actuator drive redundancy, the mode matrix is given for each possible fault. Next, based on each fault mode, the corresponding control signal is designed to ensure the stability and asymptotic tracking performance of the system, and the multi-controller set is formed. Then, the control switching mechanism is designed to select the most appropriate controller to generate control signals. Through simulations and experiments, the effectiveness of the proposed method with actuator fault compensation technology for multiple-point levitation modules of maglev vehicles was verified.

As a typical application of the low-altitude economy, UAV collaborative monitoring contributes to urban management and data collection. The dense distribution of urban buildings leads to limited perception and communication constraints, making it challenging for multi-UAV systems to achieve effective unified situational awareness. Therefore, this paper constructs a dynamic event-triggered communication strategy based on a large language model (LLM) and achieves collaborative control of multiple UAVs through deep reinforcement learning. Firstly, a LLM is used to analyze the environment code to actively understand the state of the UAV and the characteristics of the environment, effectively improving adaptability to complex environments. Then, LLM uses Python code to generate communication trigger conditions through semantic reasoning and makes dynamic adjustments to optimize communication timing and reduce network congestion and resource consumption. In addition, this paper proposes a coder-evaluator framework to solve the problems of high cost, low efficiency and intense subjectivity of the reinforcement learning human feedback (RLHF) method in LLM. In this way, the executable code generated by LLM is optimized to ensure the robustness and efficiency of the communication mechanism. Experiments in the high-fidelity AirSim simulation environment demonstrate significant improvements compared to baseline methods: a 26.58% decrease in communication cost, a 4.92% increase in task success rate, and a 21.43% acceleration in convergence speed.

High-speed maglev transportation is a new intelligent transportation system that combines high speed and eco-friendliness. The intelligent control of high-speed maglev trains (HSMT) faces many challenges such as partial actuator failure, strong nonlinearity, and input constraints ( i.e. unidirectional limitation, saturation, and dead zones). Additionally, racing effect between multiple electromagnets, external disturbances and parameter uncertainty makes it more difficult to maintain precise airgap. This paper presents a novel intelligent fault-tolerant control method to tackle the challenges of levitation control of HSMT under conditions of partial actuator failure, unidirectional inputs, saturation, and dead zones. First, a model of multi-points levitation system is provided. Then, a novel fuzzy-based control method is designed. It is based on an adaptive fuzzy update law and auxiliary manifold surfaces, which utilizes real-time measured data to estimate and compensate for the partial actuator failure and input constraints online. To the best of our knowledge, the developed fuzzy-based control law is the first method for the multi-electromagnet levitation system of HSMT that simultaneously considers partial actuator failure and input constraints. The stability and convergence within finite time of the closed-loop airgap errors are proven through the Lyapunov analysis method. Finally, experimental comparisons between the proposed fuzzy-based method and conventional control methods were conducted. The experimental results show that, compared to the LQR method, the proposed method reduces the system’s maximum overshoot by 93.3% and the maximum steady-state error by 36.98%. Moreover, it maintains stable levitation under actuator failure, which enhances the safety, efficiency, and reliability of future maglev transportation systems.

Image restoration technology is a crucial field in image processing and is extensively utilized across various domains. Recently, with advancements in graph convolutional network (GCN) technology, methods based on GCNs have increasingly been applied to image restoration, yielding impressive results. Despite these advancements, there is a gap in comprehensive research that consolidates various image denoising techniques. In this paper, we conduct a comparative study of image restoration techniques using GCNs. We begin by categorizing GCN methods into three primary application areas: image denoising, image super-resolution, and image deblurring. We then delve into the motivations and principles underlying various deep learning approaches. Subsequently, we provide both quantitative and qualitative comparisons of state-of-the-art methods using public denoising datasets. Finally, we discuss potential challenges and future directions, aiming to pave the way for further advancements in this domain. Our key findings include the identification of superior performance of GCN-based methods in capturing long-range dependencies and improving image quality across different restoration tasks, highlighting their potential for future research and applications.

The suspension system of the 12/14 bearingless switched reluctance motor (BSRM) exhibits inherent time-varying characteristics owing to the salient pole structure of its rotor. However, existing control strategies tend to overlook these timevarying characteristics, resulting in reduced suspension performances, including accuracy and resistance to interference. In response, this study explores a control strategy for addressing the time-varying nature of the suspension system. The timevarying characteristics of the suspension system are initially revealed through finite element analysis and mathematical model representation. Subsequently, a double closed-loop control system for suspension force is designed, including a time-invariant suspension force model (TISFM) and a data driven fuzzy sliding mode observer (DDFSMO). The TISFM is used to conduct a timeinvariant direct suspension force control system, whereas the DDFSMO is designed on the basis of the TISFM to compensate for the error between the TISFM and complete suspension force model (CSFM). The time-varying suspension control system is successfully converted into a time-invariant one by the combined action of the TISFM and DDFSMO. Finally, the validity of the proposed control strategy for time-varying suspension system is validated.

In the face of EMS type maglev train suspension system being susceptible to interference, non-negativity and saturation of control input, this paper adopts an adaptive robust control algorithm. By transforming the disturbance lumped term and the unmatched term in the dynamic equation, a control law is designed to compensate for parameter uncertainty and ensure input unidirectionality. The proposed method is rigorously proven to be asymptotically stable without requiring system linearization. Three sets of experiments were conducted on Changsha medium and low-speed maglev test train. The results show that the proposed method can quickly and accurately track the specified suspension gap, reduce gap oscillation and input maximum value and is robust to external interference.