检索结果-内蒙古大学图书馆

Indian Control Conference (ICC)

作者： Yogesh Dasgaonkar Radhe Shyam Sharma Raj Nayan Samant Centre for Artificial Intelligence and Robotics Indian Institute of technology Mandi India Nvidia Graphics Private Limited Mumbai Maharashtra India

ISBN: (数字)9798331517212

ISBN: (纸本)9798331517229

In a multi-robot system, two or more robots are trained to cooperate in their actions to complete a given task faster and with a higher probability of success than in a single-robot system. State-Of-The-Art (SOTA) cooperative multi-robot training methods are of two types: (a) Algorithmic: Greedy, Genetic, and (b) Machine Learning (ML): Regression, Reinforcement Learning (RL), Deep RL (DRL), and Inverse RL (IRL). Existing literature shows they perform well when the environmental conditions are well-known and static, for instance, in a car body part assembling task. To test the performance of the SOTA under dynamic conditions, we simulate a 3D fire disaster environment in a real-world urban city building with humans trapped inside it. We model the fire dynamics such that the intensity and spread rate of fire, smoke, and blocked paths increase non-linearly. The trapped humans have a high risk of losing their lives if not searched within a deadline. We state this problem as the Multi-robot Cooperative Search in Dynamic Environments (MCSDE) problem. When we applied the SOTA approaches to train multi-robots to search for human lives in this simulated environment, they failed to search for all human lives within a deadline and lacked computational efficiency. To address the issue of multi-robot training in MCSDE, we present a modular training approach. In this approach, we train each robot locally based on their local perceptions. Our efforts in this research are to uncover these perceptions' geometric locality properties. When fused with another robot's perception, each perception reveals the true behaviour of the environment's dynamic. Each robot becomes self-aware of global dynamics by training from its local perceptions. The proposed modular approach requires just one-fourth of the computation than the RL or DL-based training approaches and performs far superior to the algorithmic approaches.

关键词： Training Three-dimensional displays Heuristic algorithms Disasters Urban areas Thermal sensors Search problems Robot sensing systems Multi-robot systems Self-aware

来源：评论

学校读者我要写书评

暂无评论

BrainCLIP: Brain Representation via CLIP for Generic Natural Visual Stimulus Decoding

引用

IEEE Transactions on Medical Imaging 2025年 PP卷 PP页

作者： Ma, Yongqiang Liu, Yulong Chen, Liangjun Zhu, Guibo Chen, Badong Zheng, Nanning Xi'an Jiaotong University National Key Laboratory of Human-Machine Hybrid Augmented Intelligence National Engineering Research Center of Visual Information and Applications Institute of Artificial Intelligence and Robotics Shaanxi Xi'an710049 China Institute of Automation Beijing China

Functional Magnetic Resonance Imaging (fMRI) presents challenges due to limited paired samples and low signal-to-noise ratios, particularly in tasks involving reconstructing natural images or decoding their semantic content. To address these challenges, we introduce BrainCLIP, an innovative fMRI-based brain decoding model. BrainCLIP leverages Contrastive Language-Image Pre-training's (CLIP) cross-modal generalization abilities to bridge brain activity, images, and text for the first time. Our experiments demonstrate CLIP's effectiveness in diverse brain decoding tasks, including zero-shot visual category decoding, fMRI-image/text alignment, and fMRI-to-image generation. The core objective of BrainCLIP is to train a mapping network that translates fMRI patterns into a unified CLIP embedding space, achieved through visual and textual supervision integration. Our experiments highlight that this approach significantly enhances performance in tasks such as fMRI-text alignment and fMRI-based image generation. Notably, BrainCLIP surpasses BraVL, a recent multi-modal method, in zero-shot visual category decoding. Moreover, BrainCLIP demonstrates strong capability in reconstructing visual stimuli with high semantic fidelity, competing favorably with state-of-the-art methods in capturing high-level semantic features during fMRI-based natural image reconstruction. © 1982-2012 IEEE.

关键词： Signal to noise ratio

来源：评论

学校读者我要写书评

暂无评论

PanoOcc: Unified Occupancy Representation for Camera-based 3D Panoptic Segmentation

PanoOcc: Unified Occupancy Representation for Camera-based 3...

引用

Conference on Computer Vision and Pattern Recognition (CVPR)

作者： Yuqi Wang Yuntao Chen Xingyu Liao Lue Fan Zhaoxiang Zhang School of Artificial Intelligence University of Chinese Academy of Sciences (UCAS) CRIPAC MAIS Institute of Automation Chinese Academy of Sciences (CASIA) Centre for Artificial Intelligence and Robotics (HKISI_CAS) Shanghai AI Laboratory

ISBN: (数字)9798350353006

ISBN: (纸本)9798350353013

Comprehensive modeling of the surrounding 3D world is crucial for the success of autonomous driving. However, existing perception tasks like object detection, road structure segmentation, depth & elevation estimation, and open-set object localization each only focus on a small facet of the holistic 3D scene understanding task. This divide-and-conquer strategy simplifies the algorithm development process but comes at the cost of losing an end-to-end unified solution to the problem. In this work, we address this limitation by studying camera-based 3D panoptic segmentation, aiming to achieve a unified occupancy representation for camera-only 3D scene understanding. To achieve this, we introduce a novel method called PanoOcc, which utilizes voxel queries to aggregate spatiotemporal information from multi-frame and multi-view images in a coarse-to-fine scheme, integrating feature learning and scene representation into a unified occupancy representation. We have conducted extensive ablation studies to validate the effectiveness and efficiency of the proposed method. Our approach achieves new state-of-the-art results for camera-based semantic segmentation and panoptic segmentation on the nuScenes dataset. Furthermore, our method can be easily extended to dense occupancy prediction and has demonstrated promising performance on the Occ3D benchmark. The code will be made available at https://***/Robertwyq/PanoOcc.

关键词： Representation learning Location awareness Solid modeling Three-dimensional displays Semantic segmentation Roads Estimation

来源：评论

学校读者我要写书评

暂无评论

Zero-shot Fire And Arson Detection Using Textual Descriptions 13

Zero-shot Fire And Arson Detection Using Textual Description...

引用

13th International Conference on Information and Communication Technology Convergence, ICTC 2022

作者： Jeon, Hobeom Kim, Hyungmin Kim, Dohyung Kim, Jaehong University of Science and Technology Department of Artificial Intelligence Deajeon Korea Republic of Electronics and Telecommunications Research Institute Intelligent Robotics Research Division Deajeon Korea Republic of

ISBN: (数字)9781665499392

ISBN: (纸本)9781665499392

Recently, fire detection methods have leveraged deep learning technology to detect fire and flame texture information in images. However, training-based methods require large amounts of data to detect, and data collecting is expensive due to the fire risk. In addition, the existing dataset extracted from web videos lacks diversity in backgrounds. This paper proposes a method for detecting fire situations based on natural language descriptions to solve the data diversity problem. Our approach compares images and descriptions of the fire situation for detecting fire without training. In addition, the detector is simultaneously reasoning human arson behavior, enabling early fire detection in the surveillance videos. Our fire detection method demonstrates superior or competitive performance without training compared to other training-based methods. Furthermore, the surveillance performance of our text-based fire detector was investigated by datasets including inclement weather and nighttime situations. As a result, we confirm text-based fire detection method has high generalization performance and usability. Our approach could eventually lead to future interpretable anomaly detection studies. © 2022 IEEE.

关键词： Security systems

来源：评论

学校读者我要写书评

暂无评论

Scalable Traffic Simulation for Autonomous Driving via Multi-Agent Goal Assignment and Autoregressive Goal-Directed Planning

Scalable Traffic Simulation for Autonomous Driving via Multi...

引用

IEEE Symposium on Intelligent Vehicle

作者： Xiaoyu Mo Haochen Liu Zhiyu Huang Jianwu Fang Jianru Xue Chen Lv School of Mechanical and Aerospace Engineering Nanyang Technological University Singapore Institute of Artificial Intelligence and Robotics Xi’an Jiaotong University China

ISBN: (数字)9798350348811

ISBN: (纸本)9798350348828

Simulation provides a fast, cost-effective, and secure environment for developing autonomous driving systems. However, mitigating the gap between simulation and reality is a challenging task as it demands a behavior simulation method that is human-like, diverse, controllable, socially consistent, and scalable. This work proposes a data-driven traffic agent simulation method to address the aforementioned challenges. Our approach centers around a graph-based scene representation and an encoding method, dividing the simulation into two stages: Multi-Agent Goal assignment (MAG) and Goal-Directed Planning (GDP). Firstly, we create joint goal sets for all agents involved in the scenario. Subsequently, we assign target centerlines (TCLs) to each agent based on their predicted goals. To account for any potential mismatch between the predicted joint goal sets and the road structure, we further align the goals of each agent with their respective assigned TCLs. These on-TCL goals serve as inputs for our interactive autoregressive Goal-Directed Planner (AR-GDP), constituting the second stage of our method that generates roll-outs for simulations. Evaluation results on the leaderboard of the Waymo Open Sim Agents Challenge (WOSAC) 2023 show the competitiveness of the proposed method.

关键词： Measurement Pedestrians Economic indicators Roads Traffic control Encoding Planning

来源：评论

学校读者我要写书评

暂无评论

A Hybrid Neural Network Model for Marine Dissolved Oxygen Concentrations Time-Series Forecasting Based on Multi-Factor Analysis and a Multi-Model Ensemble

引用

Engineering 2021年第12期7卷 1751-1765页

作者： Hui Liu Rui Yang Zhu Duan Haiping Wu Institute of Artificial Intelligence and Robotics(IAIR) Key Laboratory of Traffic Safety on Track of Ministry of EducationSchool of Traffic and Transportation EngineeringCentral South UniversityChangsha 410075China

Dissolved oxygen(DO)is an important indicator of aquaculture,and its accurate forecasting can effectively improve the quality of aquatic *** this paper,a new DO hybrid forecasting model is proposed that includes three stages:multi-factor analysis,adaptive decomposition,and an optimizationbased ***,considering the complex factors affecting DO,the grey relational(GR)degree method is used to screen out the environmental factors most closely related to *** consideration of multiple factors makes model fusion more ***,the series of DO,water temperature,salinity,and oxygen saturation are decomposed adaptively into sub-series by means of the empirical wavelet transform(EWT)***,five benchmark models are utilized to forecast the sub-series of EWT *** ensemble weights of these five sub-forecasting models are calculated by particle swarm optimization and gravitational search algorithm(PSOGSA).Finally,a multi-factor ensemble model for DO is obtained by weighted *** performance of the proposed model is verified by timeseries data collected by the pacific islands ocean observing system(PacIOOS)from the WQB04 station at *** evaluation indicators involved in the experiment include the Nash–Sutcliffe efficiency(NSE),Kling–Gupta efficiency(KGE),mean absolute percent error(MAPE),standard deviation of error(SDE),and coefficient of determination(R^(2)).Example analysis demonstrates that:①The proposed model can obtain excellent DO forecasting results;②the proposed model is superior to other comparison models;and③the forecasting model can be used to analyze the trend of DO and enable managers to make better management decisions.

关键词： Dissolved oxygen concentrations forecasting Time-series multi-step forecasting Multi-factor analysis Empirical wavelet transform decomposition Multi-model optimization ensemble

来源：评论

学校读者我要写书评

暂无评论

Instant Motion Recognition of Lower Limb Using Surface Electromyography Signal Based on Convolutional Neural Network

Instant Motion Recognition of Lower Limb Using Surface Elect...

引用

2024 IEEE International Conference on robotics and Biomimetics, ROBIO 2024

作者： Li, Xiaohui Yu, Xiaoyue Zhou, Hao Jin, Dingxun Li, Guanglin Wang, Lin CAS Key Laboratory of Human-Machine Intelligence-Synergy Systems China University of Chinese Academy of Sciences China Shenzhen Lower Limb Rehabilitation Intelligent Aids Engineering Research Center China Shenzhen Technology University China Southern University of Science and Technology China SIAT Branch Shenzhen Institute of Artificial Intelligence and Robotics for Society China

ISBN: (纸本)9781665481090

The field of rehabilitation robotics has emerged as a prominent area of research in the medical community in recent years, offering innovative and promising solutions for patients with physical disabilities. The existing lower limb rehabilitation robots have some shortcomings in human-machine compatibility and motion synergy due to the limitation of the real-time human motion intention recognition. In this paper, an instant motion recognition of the lower limb was proposed using surface electromyography (sEMG) for significantly shortening the time delay between the robot motion and the actual human motion. The first 200ms data of the complete sEMG signal for a single motion cycle was employed in the recognition algorithm. The proposed recognition algorithm demonstrated superior accuracy compared to traditional machine learning methods. The accuracy of 9-class recognition was 94.32%, while 12-class recognition achieved that of 89.32%. A data augmentation method based on magnitude warping followed by shifting was employed to enhance recognition accuracy by 11.37% in 9-class recognition and 19.23% in 12-class recognition when comparing with no augmentation. The optimal results for motion recognition were observed when the first 200ms data of the complete sEMG signal for a single motion cycle with a duration of 1.0s were utilized, as evidenced by comparative results. The proposed method offers a novel approach to enhancing human-machine compatibility and comfort in the application of lower limb rehabilitation robots in the future. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

DrivingGPT: Unifying Driving World Modeling and Planning with Multi-modal Autoregressive Transformers

arXiv

引用

arXiv 2024年

作者： Chen, Yuntao Wang, Yuqi Zhang, Zhaoxiang New Laboratory of Pattern Recognition Institute of Automation Chinese Academy of Sciences China School of Artificial Intelligence University of Chinese Academy of Sciences China Centre for Artificial Intelligence and Robotics HKISI CAS Hong Kong

World model-based searching and planning are widely recognized as a promising path toward human-level physical intelligence. However, current driving world models primarily rely on video diffusion models, which specialize in visual generation but lack the flexibility to incorporate other modalities like action. In contrast, autoregressive transformers have demonstrated exceptional capability in modeling multimodal data. Our work aims to unify both driving model simulation and trajectory planning into a single sequence modeling problem. We introduce a multimodal driving language based on interleaved image and action tokens, and develop DrivingGPT to learn joint world modeling and planning through standard next-token prediction. Our DrivingGPT demonstrates strong performance in both action-conditioned video generation and end-to-end planning, outperforming strong baselines on large-scale nuPlan and NAVSIM benchmarks. Copyright © 2024, The Authors. All rights reserved.

关键词： Digital elevation model

来源：评论

学校读者我要写书评

暂无评论

Defying Imbalanced Forgetting in Class Incremental Learning

arXiv

引用

arXiv 2024年

作者： Xu, Shixiong Meng, Gaofeng Nie, Xing Ni, Bolin Fan, Bin Xiang, Shiming State Key Laboratory of Multimodal Artificial Intelligence Systems Institute of Automation Chinese Academy of Sciences China School of Artificial Intelligence University of Chinese Academy of Sciences China Centre for Artificial Intelligence and Robotics HK Institute of Science & Innovation Chinese Academy of Sciences China School of Intelligence Science and Technology University of Science and Technology Beijing China

We observe a high level of imbalance in the accuracy of different classes in the same old task for the first time. This intriguing phenomenon, discovered in replay-based Class Incremental Learning (CIL), highlights the imbalanced forgetting of learned classes, as their accuracy is similar before the occurrence of catastrophic forgetting. This discovery remains previously unidentified due to the reliance on average incremental accuracy as the measurement for CIL, which assumes that the accuracy of classes within the same task is similar. However, this assumption is invalid in the face of catastrophic forgetting. Further empirical studies indicate that this imbalanced forgetting is caused by conflicts in representation between semantically similar old and new classes. These conflicts are rooted in the data imbalance present in replay-based CIL methods. Building on these insights, we propose CLass-Aware Disentanglement (CLAD) to predict the old classes that are more likely to be forgotten and enhance their accuracy. Importantly, CLAD can be seamlessly integrated into existing CIL methods. Extensive experiments demonstrate that CLAD consistently improves current replay-based methods, resulting in performance gains of up to 2.56%. Copyright © 2024, The Authors. All rights reserved.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

Unraveling the Single Tangent Space Fallacy: An Analysis and Clarification for Applying Riemannian Geometry in Robot Learning

arXiv

引用

arXiv 2023年

作者： Jaquier, Noémie Rozo, Leonel Asfour, Tamim Institute for Anthropomatics and Robotics Karlsruhe Institute of Technology Karlsruhe Germany Bosch Center for Artificial Intelligence Renningen Germany

In the realm of robotics, numerous downstream robotics tasks leverage machine learning methods for processing, modeling, or synthesizing data. Often, this data comprises variables that inherently carry geometric constraints, such as the unit-norm condition of quaternions representing rigid-body orientations or the positive definiteness of stiffness and manipulability ellipsoids. Handling such geometric constraints effectively requires the incorporation of tools from differential geometry into the formulation of machine learning methods. In this context, Riemannian manifolds emerge as a powerful mathematical framework to handle such geometric constraints. Nevertheless, their recent adoption in robot learning has been largely characterized by a mathematically-flawed simplification, hereinafter referred to as the "single tangent space fallacy". This approach involves merely projecting the data of interest onto a single tangent (Euclidean) space, over which an off-the-shelf learning algorithm is applied. This paper provides a theoretical elucidation of various misconceptions surrounding this approach and offers experimental evidence of its shortcomings. Finally, it presents valuable insights to promote best practices when employing Riemannian geometry within robot learning applications. Copyright © 2023, The Authors. All rights reserved.

关键词： Learning algorithms

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：