Guangdong Institute of Intelligent Manufacturing

This article addresses the predefined-time optimal motion control problem of an autonomous surface vehicle (ASV) with disturbances under dynamic self-triggered frameworks via reinforcement learning (RL). Initially, to eliminate the influence of disturbance on the ASV, a predefined-time second-order integral sliding mode control (SOISM) strategy is formulated by establishing a novel integral sliding mode (ISM) function and a terminal sliding mode function. Subsequently, a predefined-time approximate optimal motion (AOM) control strategy is further developed to ensure the ASV maintains a stable state. Furthermore, a single critic network is used to obtain an approximate solution of the Hamilton-Jacobi-Bellman (HJB) equation. The above two strategies are established under the dynamic self-triggered framework, which relies on the current information to predict the next updating time, effectively reducing the computational and communication burden while avoiding the continuous monitoring of the ASV state. In the theoretical analysis, the main challenges lie in the design of Lyapunov functions and triggered conditions to ensure the stability of the sliding mode dynamics and the disturbed ASV. By applying the Lyapunov stability principle and designing two novel Lyapunov functions and triggered conditions that both contain dynamic variables, we demonstrate that the developed control strategies can ensure the stability within the specified time frame. Ultimately, simulation results verify the efficacy of the proposed motion control approach.

Compositional generalization has achieved substantial progress in computer vision on pre-collected training data. Nonetheless, real-world data continually emerges, with possible compositions being nearly infinite, long-tailed, and not entirely visible. Thus, an ideal model is supposed to gradually improve the capability of compositional generalization in an incremental manner. In this paper, we explore Composition-Incremental Learning for Compositional Generalization (CompIL) in the context of the compositional zero-shot learning (CZSL) task, where models need to continually learn new compositions, intending to improve their compositional generalization capability progressively. To quantitatively evaluate CompIL, we develop a benchmark construction pipeline leveraging existing datasets, yielding MIT-States-CompIL and C-GQA-CompIL. Furthermore, we propose a pseudo-replay framework utilizing a visual synthesizer to synthesize visual representations of learned compositions and a linguistic primitive distillation mechanism to maintain aligned primitive representations across the learning process. Extensive experiments demonstrate the effectiveness of the proposed framework.

As the maritime industry transitions towards green shipping, operational sustainability and energy efficiency are increasingly crucial for long-endurance Unmanned Surface Vehicle (USV) missions. To this end, proactively adjusting driving strategies based on the prediction of other USVs’ motion is essential. This proactive approach directly minimizes carbon emissions and reduces high-energy driving behaviors resulting from passive sudden braking or sharp turns in unexpected situations. However, existing trajectory prediction methods are trained based on low-frequency automatic identification system data of large merchant vessels, which cannot be directly used on the highly dynamic USV data. To address this limitation, this study constructs a large-scale simulated USV scenario dataset grounded in nonlinear ship hydrodynamics, which contains complicated interactive scenarios with multiple USV agents. To effectively model the interaction among agents for accurate prediction, we further propose USV-Former, a hierarchical encoder-decoder architecture designed for proactive navigation. The framework integrates a symmetric encoding structure with a dual-stage pipeline: a Local Attention Module captures high-frequency dynamics, while a Global Graph Attention Module enforces COLREGs-compliant topological constraints. Experimental results demonstrate that the proposed model outperforms established baselines in prediction accuracy. Qualitative analysis further reveals that by accurately anticipating target intentions, the model minimizes unnecessary avoidance maneuvers, enabling more stable and momentum-conserving velocity profiles. Ultimately, this architecture exhibits high computational efficiency, reduces operational energy waste, and provides a robust, measurable algorithmic foundation for green autonomous shipping and marine environmental protection.