Thrust allocation in autonomous ship berthing is a dynamic constrained multi-objective optimization problem characterized by time-varying generalized force commands, nonlinear actuator constraints, hydrodynamic interactions, and abrupt fault-induced changes in the feasible decision space. Conventional gradient-based methods are prone to local optima in nonconvex allocation spaces, whereas standard evolutionary algorithms may lack sufficient adaptability to recover high-quality Pareto solutions after constraint mutations. To address these challenges, this paper proposes a Multi-Armed Bandit-guided Evolutionary Multitasking algorithm for Dynamic Constrained Multi-Objective Optimization, termed MABEMDCMO. The original allocation problem is decomposed into a constrained main task and a relaxed auxiliary task within an evolutionary multitasking framework, enabling feasibility-oriented search while preserving objective-space diversity through inter-task knowledge transfer. A Fitness-Rate-Rank-based Multi-Armed Bandit mechanism is further developed to adaptively select between Genetic Algorithm and Differential Evolution operators according to online population-state improvement, thereby balancing exploration and exploitation without offline training or neural-network inference.In addition, a state-dependent dynamic-response strategy reconfigures exploration probability, reward-memory length, and credit decay whenever the actuator-fault or berthing-phase state indicates an active response mode, allowing operator utilities to be reassigned under the corresponding constraint regime.Based on 30 independent closed-loop CyberShip III simulations under normal operating conditions, MABEMDCMO achieves the lowest mean comprehensive metric, with an IAEW of 198509.05, mean power consumption of 6227.97 W, and mean tracking error of 31.87. Compact ablation, one-factor sensitivity analysis, and extended operator-pool diagnostics further indicate that the response-mode parameters provide robust empirical performance rather than relying on a uniquely optimal window setting. Under four representative actuator-fault scenarios, including fixed-thrust failure, thrust limitation, thrust lock-at-occurrence, and azimuth lock, the proposed algorithm maintains finite and comparatively low IAEW values, whereas several benchmark methods either suffer severe tracking degradation or fail to complete the berthing task. These results demonstrate the real-time applicability and fault-tolerant potential of MABEMDCMO for safety-critical autonomous berthing thrust allocation.

During the carrier landing process, the carrier motion induces real-time variations in the desired slope path. Considering the inherent lag in aircraft position adjustments, this paper proposes an optimal guidance law based on deep reinforcement learning (DRL) to compensate for the lag. First, the carrier landing process is modeled as a Finite Markov Decision Process (FMDP), and a comprehensive DRL framework is developed. Second, a novel Soft Actor-Critic (LA-SAC) method enhanced with the Long Short-Term Memory (LSTM) network and the attention mechanism (AM) is introduced. The method extracts the deck motion features with the LSTM network and adjusts the weights of different state data with AM to improve learning efficiency. Additionally, a distributed neural network is designed to integrate deck motion prediction and compensation, avoiding the complexity of parameter tuning in conventional methods. LA-SAC leverages full-dimensional data to train the network and derive an optimal guidance law. Finally, the superiority of the proposed method has been verified in a semi-physical simulation platform. Compared to DRL baselines, LA-SAC achieves faster convergence and derives a superior guidance policy. Compared to conventional methods, the proposed method provides a more significant lead margin to reduce landing errors. Furthermore, the ablation experiments confirmed the effectiveness of the LSTM network and AM modules, and the real-time analysis validated the practicality of the LA-SAC algorithm in actual implementation.

To improve performance for automatic carrier landing under complex wind disturbances, an active anti-disturbance control method integrating predefined-time control, disturbance observation, and online disturbance prediction is proposed. A nonlinear model carrier-based unmanned aerial vehicle (UAV) under a composite wind environment, including airwake, steady wind, and gusts, is modeled. A predefined-time sliding mode controller is then developed to ensure that the system errors converge within a user-specified time. To enhance active anti-disturbance performance, a predefined-time disturbance observer is designed for disturbance estimation, and an online prediction method based on recursive least squares with forgetting factor is introduced to predict disturbances and mitigate the lag caused by observation and UAV dynamics. Moreover, a predefined-time reference model is incorporated to avoid the exponential explosion problem. Simulation results demonstrate that, compared with the baselines, the proposed method reduces the maximum following error by 16.9–82.0% and the touchdown error by 53.4–84.1%. These results indicate that the proposed method can effectively enhance anti-disturbance performance and landing accuracy under complex wind environments.

This paper proposes an end-to-end imitation learning (IL) framework for learning artificial landing control strategies directly from human pilot demonstrations. In this setting, conventional IL methods face two key challenges, the accumulation of control errors during closed-loop execution and the inherently non-stationary and multi-modal nature of human pilot behaviors during the landing approach. To address these issues, a Multi-Intent Conditional Variational Autoencoder-based Behavior Cloning (MCVAE-BC) approach is developed. This method exploits the combined correction actions of multiple pilots during the landing glide phase by introducing a time-decaying weighted glide deviation aware dynamic time window to highlight critical correction segments. Based on this formulation, pilots' top-level correction intents and individual control styles are inferred and encoded as discrete latent representations. These latent variables are then incorporated into a behavioral cloning framework to learn artificial landing control strategies. The proposed approach is evaluated in a semi-physical carrier landing simulation environment. Experimental results demonstrate that compared to the baseline method, the MCVAE-BC framework achieves consistent pilot landing behavior and superior glide performance.

This paper investigates the automatic carrier landing control problem considering output constraints, actuator saturation, and airwake disturbances. A flexible predefined-time prescribed performance control (FPTPPC) is proposed to reconcile the conflict between output constraints and actuator saturation. First, a predefined-time prescribed performance function is designed to guarantee that the output errors converge within a user-specified time independent of initial conditions, while ensuring that both transient and steady-state performance satisfy the requirements. An error transformation is then introduced to convert the constrained domain into an equivalent unconstrained form. Subsequently, a nonsingular fast terminal integral sliding mode controller is developed to enhance the aircraft’s capability in suppressing airwake disturbances. To prevent singularities in the prescribed performance control caused by boundary violations due to actuator saturation, a flexible adjustment mechanism is embedded into the prescribed performance function (PPF). This mechanism adaptively adjusts local PPF during saturation, ensuring global landing accuracy and avoiding controller failure. Furthermore, an auxiliary system is designed to maintain stability under actuator saturation. A predefined-time reference model is intentionally introduced to eliminate the differential explosion of the command signal inherent in conventional backstepping control, and a predefined-time observer is developed to estimate and compensate for external disturbances. Finally, the effectiveness and superiority of the proposed FPTPPC have been verified through simulation experiments. Compared with dynamic inverse control, predefined-time control, and prescribed performance control, the proposed method reduces the path following errors and improves landing accuracy.

To address cooperative taxi trajectory planning for multiple aircraft on congested and constrained flight decks, trajectories must satisfy nonholonomic kinematic constraints while avoiding spatiotemporal conflicts. This paper proposes Hybrid A*-Safe Interval Path Planning (HA-SIPP), a trajectory planning method that integrates discrete-continuous hybrid validation to meet these requirements. The approach constructs a polygon-based collision model by expanding aircraft geometric contours, preserving shape characteristics while reducing unnecessary conservativeness. HA-SIPP generates kinematically feasible trajectories and resolves potential conflicts by embedding Hybrid A* motion primitives into the Safe Interval Path Planning (SIPP) framework. In particular, the hybrid validation mechanism combines safe-interval assessment in discrete state-time space with continuous-time conflict handling: conservative collision time windows are first derived using circumradii of the footprint envelopes, and a final safety certification is performed via polygon intersection checking. Experimental results on representative flight-deck layouts and fleet sizes under fixed time budgets show that HA-SIPP improves planning success and trajectory feasibility, and provides more reliable multi-aircraft conflict avoidance than traditional SIPP-based algorithms, demonstrating practical applicability in complex deck environments.

This paper proposes a robust fault-tolerant control scheme to address the challenges of automatic carrier landing systems for aircraft under air-wake disturbances, actuator faults, and output constraints. First, a predefined-time prescribed performance function (PTPPF) is constructed to explicitly confine the tracking error within a predefined range while ensuring convergence within a predefined time, independent of initial conditions. Then, a predefined-time sliding mode controller (PTSMC) is designed to enhance robustness against nonlinear uncertainties and external disturbances. Furthermore, a prescribed-time extended state observer (PTESO) is designed to accurately estimate lumped disturbances and actuator faults within a prescribed time, enabling real-time compensation in the control loop. Lyapunov-based stability analysis proves that all closed-loop signals remain bounded and the tracking error converges within the predefined-time. Carrier landing simulation experiments, considering carrier air-wake, deck motion, and actuator faults, demonstrate that the proposed method significantly improves transient performance, disturbance rejection, and landing accuracy.

This paper addresses the predefined-time landing problem for carrier aircraft under the combined effects of airwake disturbances, deck motion, and parametric uncertainties. A predefined-time control strategy is proposed, incorporating an adjustable predefined-time parameter to enable flexible tuning of the convergence time upper bound. First, a six-degree-of-freedom affine nonlinear model of the carrier aircraft is established. By integrating the direct lift control strategy, the guidance law, attitude controller, and approach power compensation subsystems are individually designed, transforming the overall tracking task into a set of subsystem stabilization problems. Second, a non-singular terminal sliding mode controller is developed based on predefined-time control theory to achieve precise landings for carrier aircraft within a predefined time. A predefined-time disturbance observer (PTDO) is also designed to estimate external disturbances, ensuring the estimation error converges to a bounded value within a predefined time. Additionally, the predefined-time reference model (PTD) is introduced to generate the derivatives of reference commands without affecting the system's global predefined-time convergence performance. The predefined-time stability of the closed-loop system is rigorously proven using Lyapunov stability theory. Finally, a series of experiments was conducted on the semi-physical simulation platform to verify the reliability and effectiveness of the proposed method.

PurposeThis study aims to evaluate the reliability of an automatic carrier landing system using predefined-time fault-tolerant control, focusing on its performance under air-wake disturbances and actuator faults.Design/methodology/approachThe research utilizes a predefined-time fault-tolerant control framework with a prescribed-time extended state observer. Simulations are conducted to assess the system's response to disturbances and faults, comparing reliability metrics like fault tolerance and convergence speed.FindingsThe results show that the predefined-time fault-tolerant control improves system reliability, offering faster state estimation and better fault-handling capabilities compared to other control methods. The system maintains stable performance under severe disturbances.Originality/valueDifferent from other control methods, this study provides a robust control solution for enhancing the safety and reliability of carrier landings, offering valuable insights for future naval aviation and aerospace control systems.

This paper proposes a surge speed control method for autonomous marine vehicles (AMVs) in the presence of fully unknown model dynamic, environmental disturbance and control input gain. By leveraging both historical and real-time input-output data, a data-driven extended state observer (DESO) is constructed for the estimation of unknown model parameters and input gain of the AMV. Then, a model-free surge control (MFSC) method is developed with relaxed data excitation by combining DESO. Considering the constrained ship-shore network communication bandwidth resources, dual-channel dynamic event-triggered mechanisms (ETMs) are incorporated into DESO and MFSC to alleviate the sampling frequency burden on the AMV. The stability analysis using Lyapunov theory demonstrates the input-to-state stability of the presented MFSC law. Simulation results validate the effectiveness of proposed dual-channel dynamic event-triggered DESO based MFSC method for AMV surge speed control within the limitation of network bandwidth resources.

Classical dense stereo matching can reconstruct ocean wave fields with high accuracy but is computationally prohibitive for real-time applications. Sparse feature tracking achieves real-time performance but produces measurements covering only 5-20% of the surface, and existing sparse-to-dense completion methods either rely on temporal sequences that fail on moving platforms, or ignore rich visual information in camera images. We propose Dual-Space Guided Propagation Network (DSGP-Net), a novel two-stage framework that introduces camera images as contextual guidance for single-frame sparse-to-dense reconstruction. The first stage processes sparse elevation data projected to image coordinates alongside camera images, directly generating a confidence-aware dense height representation in image space. The second stage refines the predictions by fusing features from both image and world coordinate systems through a cross-space fusion mechanism, maintaining geometric consistency between the two coordinate representations. Extensive experiments demonstrate that DSGP-Net achieves 14-23% RMSE reduction over state-of-the-art methods on fixed platforms and uniquely enables accurate reconstruction on moving vessels. DSGP-Net achieves real-time performance, enabling practical wave monitoring applications on autonomous surface vehicles and research vessels.

Autonomous berthing is a safety-critical operation that demands precise navigation and efficient maneuvering within complex port environments. Despite its significance, most existing research has predominantly focused on the development of low-level systems, such as control, situational awareness, and collision avoidance, while efficient path planning has received comparatively less attention. To address this gap, this study proposes a reinforcement learning (RL)-enhanced Multi-Objective Evolutionary Algorithm based on Decomposition (MOEA/D) for ship berthing path planning. The proposed approach inte -grates Deep Q-Network (DQN) and Prioritized Experience Replay (PER) to improve the evolutionary optimization process of MOEA/D. By formulating action sequences as states and employing crossover operators as actions, the method mitigates randomness and disorderly searches, thereby enhancing path planning efficiency and reducing the number of required training iterations. Additionally, the framework incorporates adaptive-length encoding, multiple crossover operators, and improved experience replay strategy to further improve berthing success rates. Experimental results demonstrate that, compared to conventional Q-Learning and DQN methods, the proposed approach generates safer berthing paths with lower time and thrust consumption within a limited training horizon.

A continuously variable transmission can improve the energy efficiency of actuators with rotary output by providing an optimum transmission ratio. A continuously variable transmission based on circumferentially arranged disks (CAD CVT) is a new type of CVT that is highly beneficial for applications requiring large torques, like heavy road transport. However, its major drawback is that its efficiency drops in the low torque region. To overcome this problem, the current paper proposes an improved mechanical design in which the force on traction disks is changed according to the instantaneous torque requirement, thus resulting in improved efficiency in low torque regions. Furthermore, a hydraulic-actuation-based control system has been designed to ensure the optimum control of the improved mechanical design. The improved mechanical design of the CAD CVT is named CAD CVT-II, which is highly beneficial for variable torque applications such as road transport and wind turbines.

All-electric ships (AESs) utilizing medium-voltage dc (MVdc) shipboard power systems (SPSs) rely on a limited number of generators to supply power to propulsion units and onboard loads. To meet the demands of dynamic maritime missions, frequent adjustments in network topology and operating power are required, which pose significant challenges for fault management by complicating fault detection and reconfiguration. This study proposes a convolutional neural network (CNN) optimized via surrogate-assisted gradient particle swarm optimization (SAGPSO) to enhance the efficiency and accuracy of short-circuit fault classification in MVdc systems. A two-zone MVdc simulation model, developed in power system computer aided design (PSCAD)/electro-magnetic transient design and control (EMTDC), is used to generate a comprehensive dataset covering high-impedance faults and various transient disturbances. These signals are converted into spectrograms using the short-time Fourier transform (STFT), enabling rich time-frequency feature representation. A supervised CNN is then employed to learn discriminative features from the spectrograms for fault classification, while SAGPSO fine-tunes the hyperparameters by incorporating gradient information and surrogate modeling to improve model performance and convergence speed. Experimental results demonstrate that the SAGPSO-CNN approach effectively distinguishes high-impedance dc interpole faults from strong transient disturbances, such as pulse load variations and sudden increases in propulsor loads, with a classification accuracy exceeding 99%. Furthermore, the method significantly refines network parameter configurations, supporting robust fault management in MVdc SPS.

With the development of deep learning, the traditional approach of manually designing network archi-tectures and determining parameters inevitably loses its reliance on expert systems, which has brought significant challenges to our parameter configuration. To address this issue, we adopted an automated search architecture, reducing manpower while enhancing efficiency. Based on this, this paper proposes the Multi-Objective Generalized Quantum Particle Swarm Optimization (MOGQPSO) algorithm, which combines quantum behavior and Gaussian distribution characteristics, introducing a brand-new mode for the movement and update of particles in the search space. Simultaneously, a bilevel optimization algorithm is designed. A particle encoding scheme is utilized to effectively represent the CNN architecture to assist in optimization. Through multiple iterations of MOGQPSO, the optimal solutions for the upper and lower levels are sought. Superior upper-level solutions are selected for lower-level optimization, and the optimized parameters of the lower level are fed back to the upper level. The upper and lower levels cooperate and iterate continuously. Moreover, combined with the decoding method in NSGA-Net, the entire CNN architecture is jointly designed. Experimental results demonstrate that the multi-objective optimal solution set obtained by this algorithm has been significantly improved compared to traditional algorithms, achieving a reduction of nearly one million parameters while enhancing accuracy.

A novel method for three-dimensional (3D) wave reconstruction based on stereo vision is proposed to overcome the challenges of measuring water surfaces under laboratory conditions. Traditional methods, such as adding seed particles or projecting artificial textures, can solve the image problem caused by the optical properties of the water surface. However, these methods can be costly and complicated to operate. In this paper, the proposed method uses affine consistency as matching invariants, bypassing the need for artificial textures. The method presents new data and smoothness terms within the graph cuts framework to achieve robust wave reconstruction. In a laboratory tank experiment, the wave point clouds were successfully reconstructed using a binocular camera. The accuracy of the method was verified by comparing the reconstruction with theoretical values and the sequences of the wave probe.

This paper aims to address the trajectory tracking problem of quadrotors under complex dynamic environments and significant fluctuations in system states. An adaptive trajectory tracking control method is proposed based on an improved Model Predictive Path Integral (MPPI) controller and a Multilayer Perceptron (MLP) neural network. The technique enhances control accuracy and robustness by adjusting control inputs in real time. The Multilayer Perceptron neural network can learn the dynamics of a quadrotor by its state parameter and then the Multilayer Perceptron sends the model to the Model Predictive Path Integral controller. The Model Predictive Path Integral controller uses the model to control the quadcopter following the desired trajectory. Experimental data show that the improved Model Predictive Path Integral–Multilayer Perceptron method reduces the trajectory tracking error by 23.7%, 34.7%, and 10.9% compared to the traditional Model Predictive Path Integral, MPC with MLP, and a two-layer network, respectively. These results demonstrate the potential application of the method in complex environments.

Among all phases of carrier-based aircraft flight, the landing process has the highest accident rate. To address safety concerns during landing, this paper proposes a novel Long Short-Term Memory model enhanced with a bilayer attention mechanism. The model utilizes flight data from carrier-based aircraft landings to predict critical touchdown states, including the position of the tail hook and the sinking speed, to determine whether the tail hook of the aircraft can safely engage with the arresting cable. By integrating a bilayer attention mechanism with the Long Short-Term Memory network, the model effectively captures temporal features and flight parameters relevant to the touchdown states. Experimental results utilizing flight data from the sliding phase demonstrate that the proposed method outperforms baseline models and delivers accurate predictions of aircraft landing touchdown states.

This paper introduces a novel attitude tracking control method for carrier-based aircraft, focusing on mitigating actuator saturation and enhancing fault tolerance. The method is fixed-time prescribed performance control with prescribed-time disturbance observer (FXTPPC-PTDO), which addresses challenges associated with carrier aircraft landings. First, the attitude motion model of aircraft carrier-based aircraft is introduced. Second, a prescribed-time disturbance observer is designed to estimate the total disturbance including actuator faults and external disturbances. Through an error transformation technique, we convert the constrained performance envelope into an equivalent unconstrained error form. Subsequently, a backstepping control strategy is proposed to ensure that this error converges to a prescribed small neighborhood within a fixed-time frame. When the attitude tracking errors must remain within performance bounds, necessitating larger actuator deflections and raising the potential for input saturation. Therefore, an anti-saturation design can improve the stability and safety of carrier-based aircraft attitude control. The simulation results of tracking the desired attitude angle show that the anti-interference, fault tolerance and tracking accuracy of FXTPPC-PTDO are better than the other control methods.