检索结果-内蒙古大学图书馆

RL + Model-Based Control: Using On-Demand Optimal Control to Learn Versatile Legged Locomotion

IEEE robotics and Automation Letters 2023年第10期8卷 6619-6626页

作者： Kang, Dongho Cheng, Jin Zamora, Miguel Zargarbashi, Fatemeh Coros, Stelian ETH Zurich Computational Robotics Lab in the Department of Computer Science Zurich8092 Switzerland

This letter presents a control framework that combines model-based optimal control and reinforcement learning (RL) to achieve versatile and robust legged locomotion. Our approach enhances the RL training process by incorporating on-demand reference motions generated through finite-horizon optimal control, covering a broad range of velocities and gaits. These reference motions serve as targets for the RL policy to imitate, leading to the development of robust control policies that can be learned with reliability. Furthermore, by utilizing realistic simulation data that captures whole-body dynamics, RL effectively overcomes the inherent limitations in reference motions imposed by modeling simplifications. We validate the robustness and controllability of the RL training process within our framework through a series of experiments. In these experiments, our method showcases its capability to generalize reference motions and effectively handle more complex locomotion tasks that may pose challenges for the simplified model, thanks to RL's flexibility. Additionally, our framework effortlessly supports the training of control policies for robots with diverse dimensions, eliminating the necessity for robot-specific adjustments in the reward function and hyperparameters. © 2016 IEEE.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Pose Estimation of an Autonomous Drilling Excavator

Pose Estimation of an Autonomous Drilling Excavator

引用

Joint International Conference of the 14th International Conference on Mechanisms and Mechanical Transmissions and 26th International Conference on robotics, MTM and robotics 2024

作者： Suiker, Marcel Husemann, Jörg Berns, Karsten Robotics Research Lab Department of Computer Science University of Kaiserslautern-Landau Kaiserslautern Germany

ISBN: (纸本)9783031875366

Due to the increase of automation in the field of construction, more and more automation approaches for excavators have been developed. A special case of excavators are drilling excavators. Therefore, a precise pose estimation is an essential task to achieve automation of the drilling process and has not been the subject of interest in research until now. This paper presents a new approach to calculating a pose of the excavator’s undercarriage for a given drilling position. The positioning uses information on the excavator’s workspace considering its kinematics, the terrain, and obstacles at the construction site. Essential to this approach is the computation of the terrain gradient, as well as the computation of the largest inscribed rectangle inside a computed candidate area. Experiments showed that the presented solution yields good results but takes a long time to find them. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Excavators

来源：评论

学校读者我要写书评

暂无评论

Deep Compliant Control for Legged Robots

Deep Compliant Control for Legged Robots

引用

IEEE International Conference on robotics and Automation (ICRA)

作者： Adrian Hartmann Dongho Kang Fatemeh Zargarbashi Miguel Zamora Stelian Coros Computational Robotics Lab in the Department of Computer Science ETH Zurich Switzerland

ISBN: (数字)9798350384574

ISBN: (纸本)9798350384581

Control policies trained using deep reinforcement learning often generate stiff, high-frequency motions in response to unexpected disturbances. To promote more natural and compliant balance recovery strategies, we propose a simple modification to the typical reinforcement learning training process. Our key insight is that stiff responses to perturbations are due to an agent’s incentive to maximize task rewards at all times, even as perturbations are being applied. As an alternative, we introduce an explicit recovery stage where tracking rewards are given irrespective of the motions generated by the control policy. This allows agents a chance to gradually recover from disturbances before attempting to carry out their main tasks. Through an in-depth analysis, we highlight both the compliant nature of the resulting control policies, as well as the benefits that compliance brings to legged locomotion. In our simulation and hardware experiments, the compliant policy achieves more robust, energy-efficient, and safe interactions with the environment.

关键词： Training Legged locomotion Uncertainty Tracking Perturbation methods Process control Energy efficiency

来源：评论

学校读者我要写书评

暂无评论

Tourist Destination Recommendation System based on Machine Learning 24

Tourist Destination Recommendation System based on Machine L...

引用

9th International Conference on Big Data and Computing, ICBDC 2024

作者： Kongpeng, Sumitra Hanskunatai, Anantaporn Data Science and Computational Intelligence Lab Department of Computer Science School of Science King Mongkut's Institute of Technology Ladkrabang Thailand

ISBN: (纸本)9798400718205

Thailand has a wide variety of tourist attractions, making it difficult for tourist to choose where to go on vacation. The tourist destination recommendation system is a challenge for creating a system to help recommend tourist destinations that are appropriate for personal. Therefore, the principal aims of this research encompass two distinct objectives: firstly, to create a recommendation system for tourist destinations in Thailand by applying machine learning algorithms;and secondly, to analyze factors influencing tourists' choices of destinations. The dataset was gathered from an online survey conducted via Google Forms, comprising responses from 429 tourists in Thailand. In the experiments, three different types of feature selection methods were applied in a data preprocessing step. In the modeling process, four machine learning algorithms, namely Decision Tree, Random Forest, k-Nearest Neighbors (k-NN), and Multi-Layer Perceptron (MLP), were used to construct the model and compare the predictive performance of the recommendation system based on hit rate and NDCG. The experimental results showed that suggesting tourist destinations in the Central region was the most effective, with the highest hit rate and NDCG compared to other regions. The average hit rate and NDCG for the five regions were 0.8 and 0.59, respectively. In addition, there has been an analysis of key factors influencing destination selection, such as activity, travel month, travel budget, and the age of tourists, to understand their impact on travel choices in each region of Thailand. © 2024 Owner/Author.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Immuno-inspired Selective Aggregation for Decentralized Federated Deep Reinforcement Learning

Immuno-inspired Selective Aggregation for Decentralized Fede...

引用

2024 Genetic and Evolutionary Computation Conference Companion, GECCO 2024 Companion

作者： Rangu, Gayathri Nair, Shivashankar Robotics Lab. Department of Computer Science and Engineering Indian Institute of Technology Guwahati Assam Guwahati India

ISBN: (纸本)9798400704956

Conventional approaches to Federated Deep Reinforcement Learning (FDRL) often mandate the participation of all the associated devices and perform indiscriminate aggregation of the models. This can, at times, culminate in a low-performance global model being pumped back to the devices causing a setback in the performances of all the local models. Aggregation should thus, be performed judiciously based on the model performance. Unlike offline Reinforcement Learning (RL) setups, when datasets and ground truths are unavailable and data needs to be gathered in real-time, defining a suitable metric for the model performance, can be challenging. In this paper, we propose a novel Immuno-inspired approach for Selective Aggregation suitable for decentralized FDRL (dFDRL) that can act as a metric for the performance of a model based on which the decision to aggregate could be made. © 2024 Copyright held by the owner/author(s).

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Pose Measurement of the EndoWrist Round Tip Scissor Instrument with Optical Coherence Tomography

Pose Measurement of the EndoWrist Round Tip Scissor Instrume...

引用

作者： Schöne, Sandra Pol, Nirmal Kahrs, Lueder A. Medical Computer Vision and Robotics Lab University of Toronto Toronto Canada University of Stuttgart Stuttgart Germany Medical Computer Vision and Robotics Lab Institute of Biomedical Engineering University of Toronto Toronto Canada Department of Mathematical and Computational Sciences University of Toronto Mississauga Mississauga Canada

To automate surgical (sub-)tasks in robotic surgery, the knowledge of the exact pose of the instrument is mandatory. The application of Optical Coherence Tomography (OCT) to the problem of pose measurement appears promising due to its advantages of 3D imaging and micron-scale resolution. To investigate this, 175 image sequences of the EndoWrist Round Tip Scissor Tool were acquired with an OCT system. The images differ in the opening angles of the scissor blades and the rotation angles of the entire instrument about its central axis. These image sequences were further processed through computer vision methods of the individual images followed by point cloud generation. For pose estimation, an Iterative Closest Point algorithm was implemented to register the acquired point clouds to reference point clouds created from the instrument CAD file. The implemented algorithm was able to determine the opening angle with an overall error of 2 ± 1.3 and the rotation angle with a standard deviation between several runs of 0.6 ±2.8. However, the overall processing time of (39 ± 17)s on a standard PC leaves room for further investigations. © 2024 by Walter de Gruyter Berlin/Boston.

关键词： Optical coherence tomography

来源：评论

学校读者我要写书评

暂无评论

Explorative Study on Motor Interference During Synchronous Human and Robot Arm Movements Under Varied Presence of a Robot Head

Explorative Study on Motor Interference During Synchronous H...

引用

2024 IEEE International Conference on robotics and Biomimetics, ROBIO 2024

作者： Kaya, Mertcan Kuhnlenz, Kolja Coburg University of Applied Sciences and Arts Robotics Research Lab Department of Electrical Engineering and Computer Science Coburg Germany

ISBN: (纸本)9781665481090

This paper investigates the influence of a static robot head on deviations of human hand movements from task direction (motor interference) during simultaneous human and robot arm movements using a collaborative robot arm. Synchronous vertical and horizontal human and robot arm movements are conducted in all combinations (within-subjects factors direction and congruency). The presence of the head (head/ no head) is varied as a between-subjects factor. In this first step, only its motionless presence is deliberately focused at, in order to exclude contributions by eye- or head-movement behavior or facial expressions being difficult variables to generalize from. Participants' hands are tracked optically marker-based and deviations from task-related vertical, respectively, horizontal movements are evaluated. In contrast to expectations, results do not show a significant impact of the robot head on motor interference. Thus, there is evidence, that appearance might not be an activating factor for motor interference in HRI. © 2024 IEEE.

关键词： Robotic arms

来源：评论

学校读者我要写书评

暂无评论

Learning in Deep Factor Graphs with Gaussian Belief Propagation 41

Learning in Deep Factor Graphs with Gaussian Belief Propagat...

引用

41st International Conference on Machine Learning, ICML 2024

作者： Nabarro, Seth van der Wilk, Mark Davison, Andrew J. Dyson Robotics Lab Imperial College London United Kingdom Department of Computer Science University of Oxford United Kingdom

We propose an approach to do learning in Gaussian factor graphs. We treat all relevant quantities (inputs, outputs, parameters, activations) as random variables in a graphical model, and view training and prediction as inference problems with different observed nodes. Our experiments show that these problems can be efficiently solved with belief propagation (BP), whose updates are inherently local, presenting exciting opportunities for distributed and asynchronous training. Our approach can be scaled to deep networks and provides a natural means to do continual learning: use the BP-estimated posterior of the current task as a prior for the next. On a video denoising task we demonstrate the benefit of learnable parameters over a classical factor graph approach and we show encouraging performance of deep factor graphs for continual image classification. Copyright 2024 by the author(s)

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

DEFVERIFY: Do Hate Speech Models Reflect Their Dataset's Definition? 31

DEFVERIFY: Do Hate Speech Models Reflect Their Dataset's Def...

引用

31st International Conference on computational Linguistics, COLING 2025

作者： Khurana, Urja Nalisnick, Eric Fokkens, Antske Computational Linguistics and Text Mining Lab Vrije Universiteit Amsterdam Netherlands Department of Computer Science Johns Hopkins University United States

ISBN: (纸本)9798891761964

When building a predictive model, it is often difficult to ensure that application-specific requirements are encoded by the model that will eventually be deployed. Consider researchers working on hate speech detection. They will have an idea of what is considered hate speech, but building a model that reflects their view accurately requires preserving those ideals throughout the workflow of data set construction and model training. Complications such as sampling bias, annotation bias, and model misspecification almost always arise, possibly resulting in a gap between the application specification and the model's actual behavior upon deployment. To address this issue for hate speech detection, we propose DEFVERIFY: a 3-step procedure that (i) encodes a user-specified definition of hate speech, (ii) quantifies to what extent the model reflects the intended definition, and (iii) tries to identify the point of failure in the workflow. We use DEFVERIFY to find gaps between definition and model behavior when applied to six popular hate speech benchmark datasets. © 2025 Association for computational Linguistics.

关键词： computational linguistics

来源：评论

学校读者我要写书评

暂无评论

The ATTUNE Model for Artificial Trust Towards Human Operators

The ATTUNE Model for Artificial Trust Towards Human Operator...

引用

2024 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2024

作者： Petousakis, Giannis Cangelosi, Angelo Stolkin, Rustam Chiou, Manolis University of Manchester Cognitive Robotics Lab Department of Computer Science United Kingdom University of Birmingham School of Metallurgy and Materials Extreme Robotics Lab United Kingdom Queen Mary University of London United Kingdom

ISBN: (纸本)9781665410205

This paper presents a novel method to quantify Trust in HRI. It proposes an HRI framework for estimating the Robot Trust towards the Human in the context of a narrow and specified task. The framework produces a real-time estimation of an AI agent's Artificial Trust towards a Human partner interacting with a mobile teleoperation robot. The approach for the framework is based on principles drawn from Theory of Mind, including information about the human state, action, and intent. The framework creates the ATTUNE model for Artificial Trust Towards Human Operators. The model uses metrics on the operator's state of attention, navigational intent, actions, and performance to quantify the Trust towards them. The model is tested on a pre-existing dataset that includes recordings (ROSbags) of a human trial in a simulated disaster response scenario. The performance of ATTUNE is evaluated through a qualitative and quantitative analysis. The results of the analyses provide insight into the next stages of the research and help refine the proposed approach. © 2024 IEEE.

关键词： Chatbots

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：