检索结果-内蒙古大学图书馆

Improving Collision-Free Success Rate For Object Goal Visual Navigation Via Two-Stage Training With Collision Prediction

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Lian, Shiwei Zhang, Feitian Robotics and Control Laboratory Department of Advanced Manufacturing and Robotics College of Engineering State Key Laboratory of Turbulence and Complex Systems Peking University Beijing100871 China

The object goal visual navigation is the task of navigating to a specific target object using egocentric visual observations. Recent end-to-end navigation models based on deep reinforcement learning have achieved remarkable performance in finding and reaching target objects. However, the collision problem of these models during navigation remains unresolved, since the collision is typically neglected when evaluating the success. Although incorporating a negative reward for collision during training appears straightforward, it results in a more conservative policy, thereby limiting the agent’s ability to reach targets. In addition, many of these models utilize only RGB observations, further increasing the difficulty of collision avoidance without depth information. To address these limitations, a new concept—collision-free success is introduced to evaluate the ability of navigation models to find a collision-free path towards the target object. A two-stage training method with collision prediction is proposed to improve the collision-free success rate of the existing navigation models using RGB observations. In the first training stage, the collision prediction module supervises the agent’s collision states during exploration to learn to predict the possible collision. In the second stage, leveraging the trained collision prediction, the agent learns to navigate to the target without collision. The experimental results in the AI2-THOR environment demonstrate that the proposed method greatly improves the collision-free success rate of different navigation models and outperforms other comparable collision-avoidance methods. Copyright © 2025, The Authors. All rights reserved.

关键词： Deep reinforcement learning

Cross-Subject Depression Level Classification Using EEG Signals with a Sample Confidence Method

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Zhang, Zhong-Yi Xu, Chen-Yang Zhao, Li-Xuan Hou, Hui-Rang Meng, Qing-Hao Tianjin Key Laboratory of Process Measurement and Control Institute of Robotics and Autonomous Systems School of Electrical and Information Engineering Tianjin University Tianjin300072 China

Electroencephalogram (EEG) is a non-invasive tool for real-time neural monitoring, widely used in depression detection via deep learning. However, existing models primarily focus on binary classification (depression/normal), lacking granularity for severity assessment. To address this, we proposed the DepL-GCN, i.e., Depression Level classification based on GCN model. This model tackles two key challenges: (1) subjectivity in depression-level labeling due to patient self-report biases, and (2) class imbalance across severity categories. Inspired by the model learning patterns, we introduced two novel modules: the sample confidence module and the minority sample penalty module. The former leverages the L2-norm of prediction errors to progressively filter EEG samples with weak label alignment during training, thereby reducing the impact of subjectivity;the latter automatically upweights misclassified minority-class samples to address imbalance issues. After testing on two public EEG datasets, DepL-GCN achieved accuracies of 81.13% and 81.36% for multi-class severity recognition, outperforming baseline models. Ablation studies confirmed both modules’ contributions. We further discussed the strengths and limitations of regression-based models for depression-level recognition. Copyright © 2025, The Authors. All rights reserved.

关键词： Electroencephalography

Unraveling tensor structures in correct-by-design controller synthesis

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Wang, Ruohan Sun, Zhiyong Haesaert, Sofie Department of Electrical Engineering Control Systems Group Eindhoven University of Technology Netherlands Department of Mechanics and Engineering Science State Key Laboratory for Turbulence and Complex Systems Peking University Beijing China

Formal safety guarantees on the synthesis of controllers for stochastic systems can be obtained using correct-by-design approaches. These approaches often use abstractions as finite-state Markov Decision Processes (MDPs). As the state space of these MDPs grows, the curse of dimensionality makes the computational and memory cost of the probabilistic guarantees, quantified with dynamic programming, scale exponentially. In this work, we leverage decoupled dynamics and unravel, via dynamic programming operations, a tree structure in the Canonical Polyadic Decomposition (CPD) of the value functions. For discrete-time stochastic systems with syntactically co-safe linear temporal logic (scLTL) specifications, we provide provable probabilistic safety guarantees and significantly alleviate the computational burden. We provide numerical studies and validations of the theoretical results on several typical case studies and showcase that the uncovered tree structure enables efficient reductions in the computational burden. Copyright © 2025, The Authors. All rights reserved.

关键词： Stochastic systems

Unit commitment with bidirectional ramping constraints of flexibility retrofit thermal power unit

学校读者我要写书评

暂无评论

International Journal of Electrical Power and Energy systems 2025年 168卷

作者： Jiang, Jingxiao Zhu, Yanran Wang, Ying Guo, Xiaorui Lv, Jianhu Qin, Jianxiang Yan, Xingyun Zhang, Kaifeng Key Laboratory of Measurement and Control of Complex Systems of Engineering Ministry of Education Southeast University Nanjing210096 China China Electric Power Research Institute Nanjing210003 China State Grid Ningxia Electric Power Co. Ltd. Yinchuan 750001 China

Flexibility retrofits of thermal power units have been widely implemented to improve power system balancing. These retrofits enable operation below the regular minimum output, known as deep peak regulation, to accommodate increasing renewable energy integration. Due to the inherent characteristics of thermal power units, ramp-up rates and ramp-down rates generally differ. Existing studies often overlook this difference by assuming equal ramping rates. To bridge this gap, this paper systematically models and analyses bidirectional ramping constraints in unit commitment (UC) models, providing a more accurate and realistic representation of the ramping process. Specifically, this paper proposes three innovative bidirectional ramping models. The first is the all-period fixed bidirectional ramping (AFBR) model, which assumes a constant ramp rate across all periods. The second model is the interperiod dynamic bidirectional ramping (InterDBR) model, which captures variations in ramping rates across different peak regulation states, including deep peak regulation with oil (DPRO), deep peak regulation (DPR), and regular peak regulation (RPR). These ramping rate adjustments are restricted to interperiod scheduling periods. The third is the intraperiod dynamic bidirectional ramping (IntraDBR) model, which considers transitions between DPRO, DPR, and RPR within an intraperiod scheduling period, thereby reflecting actual operational conditions of the unit. We validate the effectiveness of the proposed bidirectional ramping models on the IEEE 118-bus system and a real-world Ningxia Power Grid system. Results demonstrate that bidirectional ramping models more accurately and effectively support dispatch scheduling decisions, improving system security and economic efficiency. © 2025 The Authors

关键词： Retrofitting

Adapting Human Mesh Recovery with Vision-Language Feedback

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Xu, Chongyang Huang, Buzhen Zhang, Chengfang Feng, Ziliang Wang, Yangang College of Computer Science Sichuan University Chengdu610065 China Key Laboratory of Measurement and Control of Complex Systems of Engineering Ministry of Education The School of Automation Southeast University Nanjing210096 China Intelligent Policing Key Laboratory of Sichuan Province Sichuan Police College Luzhou646000 China

Human mesh recovery can be approached using either regression-based or optimization-based methods. Regression models achieve high pose accuracy but struggle with model-to-image alignment due to the lack of explicit 2D-3D correspondences. In contrast, optimization-based methods align 3D models to 2D observations but are prone to local minima and depth ambiguity. In this work, we leverage large vision-language models (VLMs) to generate interactive body part descriptions, which serve as implicit constraints to enhance 3D perception and limit the optimization space. Specifically, we formulate monocular human mesh recovery as a distribution adaptation task by integrating both 2D observations and language descriptions. To bridge the gap between text and 3D pose signals, we first train a text encoder and a pose VQ-VAE, aligning texts to body poses in a shared latent space using contrastive learning. Subsequently, we employ a diffusion-based framework to refine the initial parameters guided by gradients derived from both 2D observations and text descriptions. Finally, the model can produce poses with accurate 3D perception and image consistency. Experimental results on multiple benchmarks validate its effectiveness. The code will be made publicly available. © 2025, CC BY.

关键词： Mesh generation

Nonlinear Dynamic Modeling and Analysis of Rigid-Flexible Coupling Spacecrafts with Time-Varying Temperature Distribution

学校读者我要写书评

暂无评论

SSRN

SSRN 2025年

作者： Zhao, Bingxin Gu, Xudong Hu, Rongchun Wu, Sihan Deng, Zichen Department of Engineering Mechanics Northwestern Polytechnical University Xi’an China MIIT Key Laboratory of Dynamics and Control of Complex Systems Northwestern Polytechnical University Xi’an China

When establishing the dynamic model of a rigid-coupled spacecraft, the traditional linear method requires the flexible structure to satisfy the small deformation assumption. However, the wide range of motion of the large spacecraft may lead to nonlinear deformation of the flexible structure. In this paper, the nonlinear dynamic model of the rigid-flexible coupling spacecraft is established during in-orbit operation. Firstly, the kinetic energy of the rigid-flexible spacecraft is derived by incorporating the nonlinear deformations with second-order coupling terms. Secondly, the time-varying temperature distribution along the thickness direction of the flexible structure is derived by using a one-dimensional thermal conduction equation. Then, the nonlinear time-varying coupled dynamics equations of the rigid-flexible coupling spacecraft are obtained by the Hamilton’s variational principle. The discrete equations are derived by using the assumed modal method (AMM). Finally, the results obtained through numerical calculations indicate that the nonlinear model accurately predicts the dynamical stiffening phenomenon as the attitude rotational velocity of the rigid hub increases. At the same time, the effects of the temperature, the orbital radius and structure size on the response amplitude of the spacecraft are obtained. © 2025, The Authors. All rights reserved.

关键词： Kinetics

SmokeAgent: Multimodal agent for fine-grained smoke event analysis in large-scale wild environments

学校读者我要写书评

暂无评论

Pattern Recognition 2025年 167卷

作者： Cao, Yichao Yang, Feng Li, Xuanpeng Lu, Xiaobo School of Automation Southeast University Nanjing210096 China Key Laboratory of Measurement and Control of Complex Systems of Engineering Ministry of Education Nanjing210096 China School of Instrument Science and Engineering Southeast University Nanjing210096 China

In recent years, numerous outstanding works have emerged in the field of smoke recognition;however, challenges related to fine-grained analysis and intelligence persist in wild application scenarios. This paper approaches the problem from a new perspective, designing the first interactive Fine-Grained Smoke Analysis Agent (dubbed as SmokeAgent). Unlike traditional smoke detectors, SmokeAgent ingeniously integrates a Multi-modal Foundation Model (MFM) to detect non-rigid, semi-transparent smoke objects. It not only provides accurate fine-grained smoke category predictions but also achieves interactive close-up observations through camera manipulation. Specifically, leveraging the existing feature learning capacity of the MFM, we first design a general smoke recognition module, followed by a spatial prompt-based feature mining method to enhance the MFM's ability to analyze local smoke targets. We also design an action space for camera manipulation, allowing the agent to dynamically retrieve information and refine its fine-grained cognition in wild scenarios. Additionally, we collected the first dataset that includes both wide shot and close-up images, covering fine-grained categories such as agricultural smoke, forest fire smoke, industrial emissions, residential smoke, and fog. Extensive experiments demonstrate that the proposed SmokeAgent model offers more fine-grained and intelligent detection capabilities in wild scenarios compared to traditional smoke detectors. © 2025

关键词： Smoke detectors

A landmarks-assisted diffusion model with heatmap-guided denoising loss for high-fidelity and controllable facial image generation

学校读者我要写书评

暂无评论

Image and Vision Computing 2025年 159卷

作者： Wang, Xing Wang, Wei Su, Shixiang Lu, Mingqi Zhang, Lei Lu, Xiaobo School of Automation Southeast University Nanjing210096 China Key Laboratory of Measurement and Control of Complex Systems of Engineering Ministry of Education Nanjing210096 China Nanjing Normal University School of Electrical and Automation Engineering Nanjing210023 China

Diffusion models have significantly advanced image generation, enabling users to create diverse and realistic images from simple prompts. However, generating high-fidelity, controllable facial images remains a challenge due to the intricate details of human faces. In this paper, we present a novel diffusion model for landmarks-assisted text to face generation which directly incorporates landmarks as guidance during the diffusion process. To address the issue of global information degradation caused by fine-tuning with local information, we introduce a heatmap-guided denoising loss that selectively focuses on feature pixels most relevant to the conditioning. This biased learning strategy ensures that the model prioritizes shape and positional information, preventing excessive deterioration of its generalization ability. Unlike existing methods relying on an extra learnable branch for conditional control, our native method eliminates the conflicts inherent in dual-branch architectures when dealing with various conditions. It also enables precise manipulation of facial features, such as shape and position. Extensive experiments on CelebA-HQ and CelebAText-HQ dataset show that our method demonstrates superior performance in generating realistic and controllable facial images, outperforming existing methods in terms of fidelity, diversity, and alignment with specified landmarks. © 2025 Elsevier B.V.

关键词： Deep learning

Application of Particle Swarm Optimization Algorithm Based on Adaptive Inertia Adjustment Strategy in Scheduling of the Coking Process 26

学校读者我要写书评

暂无评论

Application of Particle Swarm Optimization Algorithm Based o...

26th International Conference on Industrial Technology, ICIT 2025

作者： Chen, Yiheng Yin, Fan Chen, Xiaochong Ren, Yi Hu, Jie Wu, Min School of Automation China University of Geosciences Wuhan430074 China Hubei Key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Wuhan430074 China Engineering Research Center of Intelligent Technology for Geo-Exploration Ministry of Education Wuhan430074 China

ISBN: (纸本)9798331521950

The sccheduling for pushing plan during the coking process critically affects the efficiency and stability of production. However, the complexity with mutiple-stage during production makes it difficult to design an efficient coke pushing plan. To address this issue, this paper proposes a scheduling method based on the particle swarm optimization algorithm. Firstly, the fuzzy c-means clustering is utilized to categorize actual operating conditions as either normal or abnormal, thereby facilitating the scheduling of pushing plan under disparate conditions. Subsequently, the scheduling problem for the pushing plan is transformed into a traveling salesman problem, and scheduling models under various conditions are established. Finally, to accelerate the convergence and enhance the algorithm's global search capability, an adaptive inertia adjustment strategy is employed to dynamically regulate the velocity and position of particles. The proposed method has been implemented in the coking process. Through the analysis of application results, the completion coefficient of pushing plan has been increased by 4.25%, demonstrating that the proposed has advantages in scheduling the pushing plan during the actual coking process. © 2025 IEEE.

关键词： Particle swarm optimization (PSO)