检索结果-内蒙古大学图书馆

Safe Multi-Agent Reinforcement Learning for Behavior-Based Cooperative Navigation

IEEE robotics and Automation Letters 2025年第6期10卷 6256-6263页

作者： Dawood, Murad Pan, Sicong Dengler, Nils Zhou, Siqi Schoellig, Angela P. Bennewitz, Maren University of Bonn Humanoid Robots Lab Bonn53113 Germany Lamarr Institute for Machine Learning and Artificial Intelligence The Center for Robotics Bonn53113 Germany Technical University of Munich Learning Systems and Robotics Lab Munchen80333 Germany

In this letter, we address the problem of behavior-based cooperative navigation of mobile robots usingsafe multi-agent reinforcement learning (MARL). Our work is the first to focus on cooperative navigation without individual reference targets for the robots, using a single target for the formation's centroid. This eliminates the complexities involved in having several path planners to control a team of robots. To ensure safety, our MARL framework uses model predictive control (MPC) to prevent actions that could lead to collisions during training and execution. We demonstrate the effectiveness of our method in simulation and on real robots, achieving safe behavior-based cooperative navigation without using individual reference targets, with zero collisions, and faster target reaching compared to baselines. Finally, we study the impact of MPC safety filters on the learning process, revealing that we achieve faster convergence during training and we show that our approach can be safely deployed on real robots, even during early stages of the training. © 2025 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Sensor Query Schedule and Sensor Noise Covariances for Accuracy-constrained Trajectory Estimation

引用

IEEE robotics and Automation Letters 2025年第7期10卷 6983-6990页

作者： Goudar, Abhishek Schoellig, Angela P. Technical University of Munich Learning Systems and Robotics Lab Germany University of Toronto Institute for Aerospace Studies Canada Vector Institute for Artificial Intelligence Canada

Trajectory estimation involves determining the trajectory of a mobile robot by combining prior knowledge about its dynamic model with noisy observations of its state obtained using sensors. The accuracy of such a procedure is dictated by the system model fidelity and the sensor parameters, such as the accuracy of the sensor (as represented by its noise covariance) and the rate at which it can generate observations, referred to as the sensor query schedule. Intuitively, high-rate measurements from accurate sensors lead to accurate trajectory estimation. However, cost and resource constraints limit the sensor accuracy and its measurement rate. Our work's novel contribution is the estimation of sensor schedules and sensor covariances necessary to achieve a specific estimation accuracy. Concretely, we focus on estimating: (i) the rate or schedule with which a sensor of known covariance must generate measurements to achieve specific estimation accuracy, and alternatively, (ii) the sensor covariance necessary to achieve specific estimation accuracy for a given sensor update rate. We formulate the problem of estimating these sensor parameters as semidefinite programs, which can be solved by off-the-shelf solvers. We validate our approach in simulation and real experiments by showing that the sensor schedules and the sensor covariances calculated using our proposed method achieve the desired trajectory estimation accuracy. Our method also identifies scenarios where certain estimation accuracy is unachievable with the given system and sensor characteristics. © 2016 IEEE.

关键词： Robot sensing systems Estimation Accuracy Schedules Trajectory Noise Time measurement Estimation error Covariance matrices Uncertainty

来源：评论

学校读者我要写书评

暂无评论

DO GENERATED DATA ALWAYS HELP CONTRASTIVE LEARNING? 12

DO GENERATED DATA ALWAYS HELP CONTRASTIVE LEARNING?

引用

12th International Conference on Learning Representations, ICLR 2024

作者： Wang, Yifei Zhang, Jizhe Wang, Yisen School of Mathematical Sciences Peking University China Institute of Artificial Intelligence and Robotics Xi'an Jiaotong University China National Key Lab of General Artificial Intelligence School of Intelligence Science and Technology Peking University China Institute for Artificial Intelligence Peking University China

Contrastive Learning (CL) has emerged as one of the most successful paradigms for unsupervised visual representation learning, yet it often depends on intensive manual data augmentations. With the rise of generative models, especially diffusion models, the ability to generate realistic images close to the real data distribution has been well recognized. These generated high-equality images have been successfully applied to enhance contrastive representation learning, a technique termed "data inflation". However, we find that the generated data (even from a good diffusion model like DDPM) may sometimes even harm contrastive learning. We investigate the causes behind this failure from the perspective of both data inflation and data augmentation. For the first time, we reveal the complementary roles that stronger data inflation should be accompanied by weaker augmentations, and vice versa. We also provide rigorous theoretical explanations for these phenomena via deriving its generalization bounds under data inflation. Drawing from these insights, we propose Adaptive Inflation (AdaInf), a purely data-centric strategy without introducing any extra computation cost. On benchmark datasets, AdaInf can bring significant improvements for various contrastive learning methods. Notably, without using external data, AdaInf obtains 94.70% linear accuracy on CIFAR-10 with SimCLR, setting a new record that surpasses many sophisticated methods. Code is available at https://***/PKU-ML/adainf. © 2024 12th International Conference on Learning Representations, ICLR 2024. All rights reserved.

关键词： Image enhancement

来源：评论

学校读者我要写书评

暂无评论

Object Stiffness Discrimination from Fingertip Forces 23

Object Stiffness Discrimination from Fingertip Forces

引用

6th International Conference on Advances in robotics, AIR 2023

作者： Dey, Abhijit Hazarika, Shyamanta Biomimetic Robotics and Artificial Intelligence Lab Indian Institute of Technology - Guwahati Assam Guwahati India

ISBN: (纸本)9781450399807

Tactile perception is crucial for detecting contact and identifying the attributes associated with it. In this research, we demonstrate that the mechanical properties of objects may be inferred by employing data-driven algorithms from fingertip contact. We collected information from a fingertip pressure sensor as it established contact with several objects using a power prehensile circular grasp. This study establishes a correlation between fingertip force and object stiffness by incorporating a clustering algorithm to assess two phases of force data for dynamic and steady state analysis. We employed the dynamic time warping (DTW) distance-based metric and the K-means clustering technique. For the two phases, we have identified three main clusters that may be characterized as pseudo labels: Soft, Medium, and Hard. To demonstrate the robustness of the model evaluation, we performed 10-fold cross-validation. For each fold, we calculated the Bouldin Index and Silhouette Coefficient and reported the mean value to evaluate the efficacy of the clusters. © 2023 ACM.

关键词： Stiffness

来源：评论

学校读者我要写书评

暂无评论

Kinematic Synthesis and Design of a Five Fingered Hand Exoskeleton 5th

Kinematic Synthesis and Design of a Five Fingered Hand Exosk...

引用

5th International and 20th National Conference on Machines and Mechanisms, iNaCoMM 2021

作者： Sarmah, Nilavjyoti Hazarika, Shyamanta M. Biomimetic Robotics and Artificial Intelligence Lab Department of Mechanical Engineering IIT Guwahati Guwahati781039 India

ISBN: (纸本)9789811937156

This paper outlines a method to design an underactuated, linkage-based hand exoskeleton.A planar 8-link mechanism with phalanges and articulations as part of the kinematic chain is used for each *** links are attached to the proximal and distal *** the thumb, a planar 6-link mechanism directly connected at the distal phalanx is *** kinematic synthesis of the mechanism based on Burmester Theory was performed with two principal considerations: *** obtain link lengths that ensured no mechanical interference between the mechanisms and the hand and *** mechanism ensured that each finger follows the natural *** mechanism is evaluated by comparing the workspace of the anatomical index finger with that of the hand exoskeleton assisted finger. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Design

来源：评论

学校读者我要写书评

暂无评论

Air2Land:A deep learning dataset for unmanned aerial vehicle autolanding from air to land

引用

IET Cyber-Systems and robotics 2022年第2期4卷 77-85页

作者： Xunchen Zheng Tianjiang Hu Machine Intelligence and Collective Robotics(MICRO)Lab Sun Yat-sen UniversityGuangzhouChina Institute for Artificial Intelligence Sun Yat-sen UniversityGuangzhouChina

In this paper,a novel deep learning dataset,called Air2Land,is presented for advancing the state‐of‐the‐art object detection and pose estimation in the context of one fixed‐wing unmanned aerial vehicle autolanding *** bridges vision and control for ground‐based vision guidance systems having the multi‐modal data obtained by diverse sensors and pushes forward the development of computer vision and autopilot algorithms tar-geted at visually assisted landing of one fixed‐wing *** dataset is composed of sequential stereo images and synchronised sensor data,in terms of the flying vehicle pose and Pan‐Tilt Unit angles,simulated in various climate conditions and landing *** real‐world automated landing data is very limited,the proposed dataset provides the necessary foundation for vision‐based tasks such as flying vehicle detection,key point localisation,pose estimation ***,in addition to providing plentiful and scene‐rich data,the developed dataset covers high‐risk scenarios that are hardly accessible in *** dataset is also open and available at https://***/micros‐uav/micros_air2land as well.

关键词： autonomous landing(autolanding) deep learning,ground stereo vision open dataset unmanned aerial vehicle(UAV)

来源：评论

学校读者我要写书评

暂无评论

Integrative Analysis of Movement: AI-Enhanced Video and Inertial Sensors in Athletic Contexts

Integrative Analysis of Movement: AI-Enhanced Video and Iner...

引用

International Workshop on Engineering Methodologies for Medicine and Sports, EMMS 2024

作者： Zanela, Andrea Caprioli, Lucio Frontuto, Cristiana Bonaiuto, Vincenzo Robotics and Artificial Intelligence Lab ENEA "Casaccia" Research Centre Rome Italy Sports Engineering Lab Department Industrial Engineering University of Rome Tor Vergata Rome Italy

ISBN: (纸本)9783031637544

An in-depth examination of human movement proves invaluable for healthcare professionals and trainers, facilitating targeted interventions for specific pathologies, assessing movement compensation in post-injury activity, or identifying particular technical flaws. The exploited video analysis system excels in extracting pivotal features associated with subjects’ movements, and it has been successfully used in diverse scenarios, encompassing clinical contexts with neurodegenerative disease patients and sports environments with both amateur and professional athletes. Wearable inertial devices can provide comparable information, although they may involve, due to their invasiveness, a possible alteration of the specific movement. Nevertheless, they are very useful when video analysis cannot be used due to occlusions or other recording difficulties. This preliminary study aims to integrate movement measurements obtained through a marker less video analysis system that leverages artificial intelligence techniques with those acquired through inertial sensors directly placed on subjects. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

关键词： Neurodegenerative diseases

来源：评论

学校读者我要写书评

暂无评论

Optimization-inspired manual architecture design and neural architecture search

引用

Science China(Information Sciences) 2023年第11期66卷 100-112页

作者： Yibo YANG Zhengyang SHEN Huan LI Zhouchen LIN JD Explore Academy School of Mathematical Sciences Peking University Institute of Robotics and Automatic Information Systems College of Artificial Intelligence Nankai University Key Lab of General Artificial Intelligence School of Intelligence Science and Technology Peking University Institute for Artificial Intelligence Peking University Pazhou Lab

Neural architecture has been a research focus in recent years due to its importance in deciding the performance of deep networks. Representative ones include a residual network(Res Net) with skip connections and a dense network(Dense Net) with dense connections. However, a theoretical guidance for manual architecture design and neural architecture search(NAS) is still lacking. In this paper, we propose a manual architecture design framework, which is inspired by optimization algorithms. It is based on the conjecture that an optimization algorithm with a good convergence rate may imply a neural architecture with good performance. Concretely, we prove under certain conditions that forward propagation in a deep neural network is equivalent to the iterative optimization procedure of the gradient descent algorithm minimizing a cost function. Inspired by this correspondence, we derive neural architectures from fast optimization algorithms,including the heavy ball algorithm and Nesterov's accelerated gradient descent algorithm. Surprisingly, we find that we can deem the Res Net and Dense Net as special cases of the optimization-inspired *** architectures offer not only theoretical guidance, but also good performances in image recognition on multiple datasets, including CIFAR-10, CIFAR-100, and Image Net. Moreover, we show that our method is also useful for NAS by offering a good initial search point or guiding the search space.

关键词： deep neural network manual architecture design neural architecture search image recognition optimization algorithms learning-based optimization

来源：评论

学校读者我要写书评

暂无评论

Proportion integral-type active disturbance rejection generalized predictive control for distillation process based on grey wolf optimization parameter tuning

引用

Chinese Journal of Chemical Engineering 2022年第9期35卷 234-244页

作者： Jia Ren Zengqiang Chen Mingwei Sun Qinglin Sun Zenghui Wang College of Artificial Intelligence Nankai UniversityTianjin 300350China Key Lab of Intelligent Robotics of Tianjin Tianjin 300350China Department of Electrical and Mining Engineering University of South AfricaFlorida 1710South Africa

The high-purity distillation column system is strongly nonlinear and coupled,which makes it difficult to *** disturbance rejection control(ADRC)has been widely used in distillation systems,but it has limitations in controlling distillation systems with large time delays since ADRC employs ESO and feedback control law to estimate the total disturbance of the system without considering the large time *** paper designs a proportion integral-type active disturbance rejection generalized predictive control(PI-ADRGPC)algorithm to control the distillation column system with large time *** replaces the PD controller in ADRC with a proportion integral-type generalized predictive control(PI-GPC),thereby improving the performance of control systems with large time *** the proposed controller has many parameters and is difficult to tune,this paper proposes to use the grey wolf optimization(GWO)to tune these parameters,whose structure can also be used by other intelligent optimization *** performance of GWO tuned PI-ADRGPC is compared with the control performance of GWO tuned ADRC method,multi-verse optimizer(MVO)tuned PI-ADRGPC and MVO tuned *** simulation results show that the proposed strategy can track reference well and has a good disturbance rejection performance.

关键词： Proportion integral-type active disturbance rejection generalized predictive control Grey wolf optimization Parameter tuning Distillation Process control Prediction

来源：评论

学校读者我要写书评

暂无评论

Voice Command Drone for Law Enforcement and Emergency Response 9

Voice Command Drone for Law Enforcement and Emergency Respon...

引用

9th International Conference on Mobile and Secure Services, MOBISECSERV 2024

作者： Majumdar, Swarnamouli Kirkley, Sonny Srivastava, Mayur Zenext Ai Artificial Intelligence & Robotics Montréal Canada Indiana University Crisis Technologies Innovation Lab BloomingtonIN United States Zenext Ai Design & Ux AkronOH United States

ISBN: (纸本)9798350390896

In an era where the rapid response and situational awareness of law enforcement and emergency responders can mean the difference between life and death, leveraging cutting-edge technology is paramount. This paper presents a compelling proposal for integrating Zenext AI's advanced voice-command technology with unmanned aerial systems (UAS) to revolutionize operations in law enforcement and emergency response. These voice-controlled drones will perform critical tasks such as registration plate checks, image capture, suspect tracking, disaster rescue missions, and firefighting support, all initiated through simple voice commands. This innovative solution aims to enhance the operational efficiency, safety, and situational awareness of police officers, firefighters, and other first responders, ensuring they have the best tools to protect and serve the community effectively. © 2024 IEEE.

关键词： Fires

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：