检索结果-内蒙古大学图书馆

LiDAR-based Registration against Georeferenced Models for Globally Consistent Allocentric Maps

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Quenzel, Jan Mallwitz, Linus T. Arnold, Benedikt T. Behnke, Sven Autonomous Intelligent Systems Group Computer Science Institute VI – Intelligent Systems and Robotics Germany Center for Robotics Lamarr Institute for Machine Learning and Artificial Intelligence University of Bonn Germany Fraunhofer FIT Germany Fraunhofer IAIS Germany

Modern unmanned aerial vehicles (UAVs) are irreplaceable in search and rescue (SAR) missions to obtain a situational overview or provide closeups without endangering personnel. However, UAVs heavily rely on global navigation satellite system (GNSS) for localization which works well in open spaces, but the precision drastically degrades in the vicinity of buildings. These inaccuracies hinder aggregation of diverse data from multiple sources in a unified georeferenced frame for SAR operators. In contrast, CityGML models provide approximate building shapes with accurate georeferenced poses. Besides, LiDAR works best in the vicinity of 3D structures. Hence, we refine coarse GNSS measurements by registering LiDAR maps against CityGML and digital elevation map (DEM) models as a prior for allocentric mapping. An intuitive plausibility score selects the best hypothesis based on occupancy using a 2D height map. Afterwards, we integrate the registration results in a continuous-time spline-based pose graph optimizer with LiDAR odometry and further sensing modalities to obtain globally consistent, georeferenced trajectories and maps. We evaluate the viability of our approach on multiple flights captured at two distinct testing sites. Our method successfully reduced GNSS offset errors from up-to 16 m to below 0.5 m on multiple flights. Furthermore, we obtain globally consistent maps w.r.t. prior 3D geospatial models. Copyright © 2024, The Authors. All rights reserved.

关键词： Continuous time systems

Novel Dynamic Feature Fusion Stragegy for Detection of Small Underwater Marine Object 5

学校读者我要写书评

暂无评论

Novel Dynamic Feature Fusion Stragegy for Detection of Small...

5th International Conference on Intelligent Autonomous Systems, ICoIAS 2022

作者： Chen, Jie Joo, Meng Zhang, Yani Gao, Wenxiao Wu, Jianguo Institute of Artificial Intelligence and Marine Robotics College of Marine Electrical Engineering Dalian Maritime University Dalian116026 China Tianjin Hanhai Lanfan Marine Technology Co. Ltd China

ISBN: (数字)9781665498388

ISBN: (纸本)9781665498388

Underwater marine object detection has been shown to be significant in the community of marine science and engineering, and has been extensively employed in many practical applications, including commercial fisheries, marine exploration, etc. A typical small object detection trap frequently exists due to the fact that the majority of objects of interest in the underwater environment are usually small. Detection of small marine object is highly desired. This paper presents a novel dynamic feature fusion algorithm based on spatial-Aware and scale-Aware attention that adaptively fuses different scale feature maps for the detection of small underwater marine objects. Leveraging the proposed dynamic feature fusion algorithm, an improved YOLOX-DFF model is proposed to detect marine organisms in natural seabed environments. To alleviate conflicts between classification and regression tasks in the detection head, the decoupled head of YOLOX-DFF is also augmented with a task-Aware attention module. The proposed model surpasses state-of-The-Art methods by a large margin of +0.7mAP, and +2.3AP{{S}} on the DUO dataset. The effectiveness and efficiency of each module are further demonstrated by the ablation studies on PASCAL VOC 0712 datasets, achieving excellent performance with +1.3mAP, and +3.0AP{S} improvements. © 2022 IEEE.

关键词： Object detection

A Review on On Device Privacy and Machine Learning Training 2

学校读者我要写书评

暂无评论

A Review on On Device Privacy and Machine Learning Training

2nd IEEE International Conference on artificial intelligence in Everything, AIE 2022

作者： Tonga, Paul Audu Ameen, Zubaida Said Mubarak, Auwalu Saleh Al-Turjman, Fadi Ai and Robotics Institute Near East University Research Center for Ai and IoT Artificial Intelligence Dept Turkey University of Kyrenia Research Center for Ai and IoT Faculty of Engineering Kyrenia Cyprus

ISBN: (数字)9781665474009

ISBN: (纸本)9781665474009

On device machine learning has in recent times attracted growing attention and applications, ranging from advanced OCR and handwriting recognition, smart sleep tracking in Apple iOS, mobile virtual assistants and Image recognition, text translation in android;and more. As a result, there is increasing demand for user privacy and security in machine learning applications. Federated Learning (FL), a new method, allows users to cooperatively train a global model while keeping the data local, thereby resolving the issue of data privacy protection through the encryption mechanism. The server aggregates models until convergence while the clients train their local models. In this procedure, the server employs a reward system to persuade users to provide huge amounts of high-quality data in order to enhance the overall model. This paper explores a few FL techniques that can be applied alongside on-device learning, in order to achieve better privacy and security. © 2022 IEEE.

关键词： Internet of things

Imagine, Initialize, and Explore: An Effective Exploration Method in Multi-Agent Reinforcement Learning

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Liu, Zeyang Wan, Lipeng Yang, Xinrui Chen, Zhuoran Chen, Xingyu Lan, Xuguang National Key Laboratory of Human-Machine Hybrid Augmented Intelligence National Engineering Research Center for Visual Information and Application Institute of Artificial Intelligence and Robotics Xi’an Jiaotong University Xi’an710049 China

Effective exploration is crucial to discovering optimal strategies for multi-agent reinforcement learning (MARL) in complex coordination tasks. Existing methods mainly utilize intrinsic rewards to enable committed exploration or use role-based learning for decomposing joint action spaces instead of directly conducting a collective search in the entire action-observation space. However, they often face challenges obtaining specific joint action sequences to reach successful states in long-horizon tasks. To address this limitation, we propose Imagine, Initialize, and Explore (IIE), a novel method that offers a promising solution for efficient multi-agent exploration in complex scenarios. IIE employs a transformer model to imagine how the agents reach a critical state that can influence each other’s transition functions. Then, we initialize the environment at this state using a simulator before the exploration phase. We formulate the imagination as a sequence modeling problem, where the states, observations, prompts, actions, and rewards are predicted autoregressively. The prompt consists of timestep-to-go, return-to-go, influence value, and one-shot demonstration, specifying the desired state and trajectory as well as guiding the action generation. By initializing agents at the critical states, IIE significantly increases the likelihood of discovering potentially important under-explored regions. Despite its simplicity, empirical results demonstrate that our method outperforms multi-agent exploration baselines on the StarCraft Multi-Agent Challenge (SMAC) and SMACv2 environments. Particularly, IIE shows improved performance in the sparse-reward SMAC tasks and produces more effective curricula over the initialized states than other generative methods, such as CVAE-GAN and diffusion models. © 2024, CC BY.

关键词： Reinforcement learning

Siren: A Learning-Based Multi-Turn Attack Framework for Simulating Real-World Human Jailbreak Behaviors

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Zhao, Yi Zhang, Youzhi Department of Computing The Hong Kong Polytechnic University Hong Kong Centre for Artificial Intelligence and Robotics Hong Kong Institute of Science & Innovation Chinese Academy of Sciences China

Large language models (LLMs) are widely used in real-world applications, raising concerns about their safety and trustworthiness. While red-teaming with jailbreak prompts exposes the vulnerabilities of LLMs, current efforts focus primarily on single-turn attacks, overlooking the multi-turn strategies used by real-world adversaries. Existing multi-turn methods rely on static patterns or predefined logical chains, failing to account for the dynamic strategies during attacks. We propose Siren, a learning-based multi-turn attack framework designed to simulate real-world human jailbreak behaviors. Siren consists of three stages: (1) training set construction utilizing Turn-Level LLM feedback (Turn-MF), (2) post-training attackers with supervised fine-tuning (SFT) and direct preference optimization (DPO), and (3) interactions between the attacking and target LLMs. Experiments demonstrate that Siren achieves an attack success rate (ASR) of 90% with LLaMA-3-8B as the attacker against Gemini-1.5Pro as the target model, and 70% with Mistral-7B against GPT-4o, significantly outperforming single-turn baselines. Moreover, Siren with a 7B-scale model achieves performance comparable to a multi-turn baseline that leverages GPT-4o as the attacker, while requiring fewer turns and employing decomposition strategies that are better semantically aligned with attack goals. We hope Siren inspires the development of stronger defenses against advanced multi-turn jailbreak attacks under realistic scenarios. Code is available at https://***/YiyiyiZhao/siren. Warning: This paper contains potentially harmful text. © 2025, CC BY-NC-SA.

关键词： Semantics

AI-based Framework for Robust Model-Based Connector Mating in Robotic Wire Harness Installation

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Kienle, Claudius Alt, Benjamin Schneider, Finn Pertlwieser, Tobias Jäkel, Rainer Rayyes, Rania ArtiMinds Robotics Karlsruhe Germany IAS Lab Computer Science Department TU Darmstadt Germany AICOR Institute for Artificial Intelligence University of Bremen Germany Karlsruhe Germany

Despite the widespread adoption of industrial robots in automotive assembly, wire harness installation remains a largely manual process, as it requires precise and flexible manipulation. To address this challenge, we design a novel AI-based framework that automates cable connector mating by integrating force control with deep visuotactile learning. Our system optimizes search-and-insertion strategies using first-order optimization over a multimodal transformer architecture trained on visual, tactile, and proprioceptive data. Additionally, we design a novel automated data collection and optimization pipeline that minimizes the need for machine learning expertise. The framework optimizes robot programs that run natively on standard industrial controllers, permitting human experts to audit and certify them. Experimental validations on a center console assembly task demonstrate significant improvements in cycle times and robustness compared to conventional robot programming approaches. Videos are available under https: //***/*** Codes 68T40 © 2025, CC BY.

关键词： Robot programming

Optimal scheduling of islanded microgrid based on improved MOSSA 43

学校读者我要写书评

暂无评论

Optimal scheduling of islanded microgrid based on improved M...

43rd Chinese Control Conference, CCC 2024

作者： Huanhuan, Lai Qingqi, Pei Meiqin, Liu Shanling, Dong Zhiyun, Xie Jian, Zheng Hezuo, Qu Senlin, Zhang Wenzhou Branch of State Grid Wenzhou325028 China Zhejiang University College of Electrical Engineering Hangzhou310027 China Zhejiang University State Key Laboratory of Industrial Control Technology Hangzhou310027 China Xi'an Jiaotong University Institute of Artificial Intelligence and Robotics Xi'an710049 China Jinhua Research Institute of Zhejiang University Jinhua321036 China

ISBN: (纸本)9789887581581

Microgrid systems are widely used in various scenarios. However, there is great uncertainty in the output of wind and solar units. How to maximize both the economic and environmental benefits of microgrid is an important issue. To achieve this, considering consumption of renewable energy and total cost, a multi-objective optimization model for islanded microgrid is established. The classical optimization algorithm has insufficient ability to approach the optimal solution in each dimension when facing multi-objective optimization. Thus, this paper introduces the fuzzy clustering mechanism and the non-dominated sorting method to obtain the improved multi-objective sparrow search algorithm (MOSSA). Taking data of renewable energy units and loads from a certain microgrid as test case, the Pareto frontier is searched by improved MOSSA. The experimental results show that the proposed method can effectively reduce the economic cost and improve the absorption capacity of wind power and photovoltaic energy. © 2024 Technical Committee on Control Theory, Chinese Association of Automation.

关键词： improved MOSSA microgrid multi-objective optimization Pareto frontier

Robust Noisy Label Learning via Two-Stream Sample Distillation

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Bai, Sihan Zhou, Sanping Qin, Zheng Wang, Le Zheng, Nanning The National Key Laboratory of Human-Machine Hybrid Augmented Intelligence National Engineering Research Center for Visual Information and Applications Institute of Artificial Intelligence and Robotics Xi’an Jiaotong University Shaanxi 710049 China

Noisy label learning aims to learn robust networks under the supervision of noisy labels, which plays a critical role in deep learning. Existing work either conducts s ample selection or label correction to deal with noisy labels during the model training process. In this paper, we design a simple yet effective sample selection framework, termed Two-Stream Sample Distillation (TSSD), for noisy label learning, which can extract more high-quality samples with clean labels to improve the robustness of network training. Firstly, a novel Parallel Sample Division (PSD) module is designed to generate a certain training set with sufficient reliable positive and negative samples by jointly considering the sample structure in feature space and the human prior in loss space. Secondly, a novel Meta Sample Purification (MSP) module is further designed to mine adequate semi-hard samples from the remaining uncertain training set by learning a strong meta classifier with extra golden data. As a result, more and more high-quality samples will be distilled from the noisy training set to train networks robustly in every iteration. Extensive experiments on four benchmark datasets, including CIFAR-10, CIFAR-100, Tiny-ImageNet, and Clothing-1M, show that our method has achieved state-of-the-art results over its competitors. Copyright © 2024, The Authors. All rights reserved.

关键词： Image classification

On the O(T√1/d4) Convergence Rate of RMSProp and Its Momentum Extension Measured by 1 Norm

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Li, Huan Dong, Yiming Lin, Zhouchen Institute of Robotics and Automatic Information Systems College of Artificial Intelligence Nankai University Tianjin China National Key Lab of General AI School of Intelligence Science and Technology Peking University Beijing China

Although adaptive gradient methods have been extensively used in deep learning, their convergence rates proved in the literature are all slower than that of SGD, particularly with respect to their dependence on the dimension. This paper considers the classical RMSProp and its momentum extension and establishes the convergence rate of T1 PTk=1 E [∇f(xk)1] ≤ O(√T1dC/4 ) measured by 1 norm without the bounded gradient assumption, where d is the dimension of the optimization variable, T is the iteration number, and C is a constant identical to that appeared in the optimal convergence rate of SGD. Our convergence rate matches the lower bound with respect to all the coefficients except the dimension d. Since x2 x1 ≤ √dx2 for problems with extremely large d, our convergence rate can be considered to be analogous to the T1 PTk=1 E [∇f(xk)2] ≤ O(TC1/4 ) rate of SGD in the ideal case of ∇f(x)1 = Θ(√d)∇f(x)2 © 2024, CC BY.

关键词： Gradient methods