检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Zhao, Wenda Goudar, Abhishek Tang, Mingliang Qiao, Xinyuan Schoellig, Angela P. Institute for Aerospace Studies University of Toronto The University of Toronto Robotics Institute Vector Institute for Artificial Intelligence Toronto Canada Munich Germany

Ultra-wideband (UWB) time difference of arrival (TDOA)-based localization has emerged as a low-cost and scalable indoor positioning solution. However, in cluttered environments, the performance of UWB TDOA-based localization deteriorates due to the biased and non-Gaussian noise distributions induced by obstacles. In this work, we present a bi-level optimization-based joint localization and noise model learning algorithm to address this problem. In particular, we use a Gaussian mixture model (GMM) to approximate the measurement noise distribution. We explicitly incorporate the estimated state's uncertainty into the GMM noise model learning, referred to as uncertainty-aware GMM, to improve both noise modeling and localization performance. We first evaluate the GMM noise model learning and localization performance in numerous simulation scenarios. We then demonstrate the effectiveness of our algorithm in extensive real-world experiments using two different cluttered environments. We show that our algorithm provides accurate position estimates with low-cost UWB sensors, no prior knowledge about the obstacles in the space, and a significant amount of UWB radios occluded. Copyright © 2023, The Authors. All rights reserved.

关键词： Uncertainty analysis

来源：评论

学校读者我要写书评

暂无评论

Vision-Based Autonomous Tracking of High-Rise Vertical Structure Using a Quadcopter

Vision-Based Autonomous Tracking of High-Rise Vertical Struc...

引用

International Conference on Computer and Automation Engineering, ICCAE

作者： Ayush Gupta Amit Shukla Ashok Kumar Sivarathri Naisarg Pandya School of Mechanical and Materials Engineering Indian Institute of Technology Mandi Mandi India Centre for Artificial Intelligence and Robotics Indian Institute of Technology Mandi Mandi India

ISBN: (数字)9798350370058

ISBN: (纸本)9798350370164

Inspection of the high-rise vertical structure is one of the most important and difficult tasks in industries for continuous operation. The majority of industrial subsurface utilities are manually examined, which takes a lot of time, money, and planning. Robot-based autonomous navigation and inspection can be a potential solution to this problem. An autonomous drone that can autonomously navigate around the structure at a high speed and has a significant payload capacity can improve the inspection quality and can perform at the minimum time with a high frequency of inspection. This research paper focuses on developing vision-based autonomous navigation and motion control algorithms for quadcopters to track high-rise industrial vertical structures. For the detection of the vertical structure, we have employed a CNN-based image segmentation model, and for the motion control for navigation around the structure, a linear control algorithm has been used. This paper discusses all the results related to the positioning and motion control of the drone, the robot vision algorithm, and the two distinct paths that the quadcopter has followed.

关键词： Image segmentation Computer vision Tracking Navigation Inspection Planning Motion control

来源：评论

学校读者我要写书评

暂无评论

Robust multiple source localization in low SNR environment based on overdetermined independent component analysis in spherical harmonic domain 24

Robust multiple source localization in low SNR environment b...

引用

24th International Congress on Acoustics, ICA 2022

作者： Hu, Meiling Hu, Yuxiang Lu, Jing Key Laboratory of Modern Acoustics Nanjing University Nanjing210093 China Horizon Robotics Beijing100094 China Nanjing Institute of Advanced Artificial Intelligence Nanjing210014 China

Spherical microphone array is often utilized to estimate the direction-of-arrival (DOA) of sound sources due to its symmetric configuration and scalability. Independent component analysis (ICA) in the spherical harmonics (SH) domain is an appealing estimation method since theoretically all time-frequency (TF) bins can be efficiently processed in the same formulation. Current ICA-based DOA estimation in SH domain assumes that the number of the sources is the same as the order of the SH expansion, and does not take noise into account. In this paper, we propose a DOA estimation method for multiple sources based on overdetermined ICA in SH domain. The principal component analysis (PCA) is utilized to transfer the overdetermined model into a determined one, and a regularization technique is further applied to alleviate the influence of noise. The robustness of the proposed method is validated by simulations. © 2022 Proceedings of the International Congress on Acoustics. All rights reserved.

关键词： Direction of arrival

来源：评论

学校读者我要写书评

暂无评论

Model Predictive Control for Nonlinear Bilateral Teleoperation Systems With Time Delay 41

Model Predictive Control for Nonlinear Bilateral Teleoperati...

引用

第41届中国控制会议

作者： Yu Liu Chuanzhi Zang Yuqi Liu Peng Zeng School of Artificial Intelligence Shenyang University of Technology Key Laboratory of Networked Control Systems Chinese Academy of Sciences Shenyang Institute of Automation Chinese Academy of Sciences Institutes for Robotics and Intelligent Manufacturing Chinese Academy of Sciences State Key Laboratory of Robotics Shenyang Institute of AutomationChinese Academy of Sciences

ISBN: (数字)9789887581536

ISBN: (纸本)9781665482561

Bilateral teleoperation systems have received extensive attention as a substitute for the human to perform tasks in remote and hazardous *** paper proposes a model predictive control strategy for nonlinear bilateral teleoperation systems based on Long-Short Term Memory(LSTM) ***,model predictive control is used to achieve good tracking performance in the presence of input constraints in the *** the same time,considering the influence of different proficiency operators operating the master manipulator on the system performance,the error between the actual master manipulator trajectory and the reference trajectory is introduced into a linear feedback term to compensate for the decrease in system control accuracy caused by the error of ***,the LSTM network is used to predict the trajectory of the master manipulator in the future time equivalent to the network transmission *** predictions are used to control the remote slave manipulator,thereby eliminating the impact of communication delay on the ***,in slave sides,the trajectory creators are applied to generate the desired trajectories to achieve good force feedback *** experiments are carried out to verify the proposed control strategy,which can guarantee good performance with both position tracking and force feedback under time delay.

关键词： Model predictive control bilateral teleoperation long-short term memory network communication delay

来源：评论

学校读者我要写书评

暂无评论

Path Following Control of Underactuated AUV Based on Improved Model Predictive Control

Path Following Control of Underactuated AUV Based on Improve...

引用

International Conference on Intelligent Autonomous Systems (ICoIAS)

作者： Siqi Wang Meng Joo Er Tianhe Liu Huibin Gong Institute of Artificial Intelligence and Marine Robotics College of Marine Electrical Engineering Dalian Maritime University Dalian China

In this paper, an improved model predictive control (MPC) method using line-of-sight (LOS) for precise path following control is proposed. First, simplify the motion model of the underactuation five degree of freedom (DOF) autonomous underwater vehicle (AUV), Yunfan-04. The LOS guidance method provides precise heading instructions based on the line of sight position between the AUV and target. The MPC algorithm is used to generate an optimal control strategy based on the dynamic model of AUV to control the following motion of AUV. In order to optimize the path following performance, the problem formula is transformed into an optimization problem with constraints, and the optimal control input is solved. Simulation experiments validate our approach. The LOS and MPC-based method outper-forms traditional methods in accuracy and robustness, achieving stable and precise path following. This work provides valuable insights for developing AUV control systems.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Grounded answers for multi-agent decision-making problem through generative world model 24

Grounded answers for multi-agent decision-making problem thr...

引用

Proceedings of the 38th International Conference on Neural Information Processing Systems

作者： Zeyang Liu Xinrui Yang Shiguang Sun Long Qian Lipeng Wan Xingyu Chen Xuguang Lan 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'an China

ISBN: (纸本)9798331314385

Recent progress in generative models has stimulated significant innovations in many fields, such as image generation and chatbots. Despite their success, these models often produce sketchy and misleading solutions for complex multi-agent decisionmaking problems because they miss the trial-and-error experience and reasoning as humans. To address this limitation, we explore a paradigm that integrates a language-guided simulator into the multi-agent reinforcement learning pipeline to enhance the generated answer. The simulator is a world model that separately learns dynamics and reward, where the dynamics model comprises an image tokenizer as well as a causal transformer to generate interaction transitions autoregressively, and the reward model is a bidirectional transformer learned by maximizing the likelihood of trajectories in the expert demonstrations under language guidance. Given an image of the current state and the task description, we use the world model to train the joint policy and produce the image sequence as the answer by running the converged policy on the dynamics model. The empirical results demonstrate that this framework can improve the answers for multi-agent decision-making problems by showing superior performance on the training and unseen tasks of the StarCraft Multi-Agent Challenge benchmark. In particular, it can generate consistent interaction sequences and explainable reward functions at interaction states, opening the path for training generative models of the future.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Registered and Segmented Deformable Object Reconstruction from a Single View Point Cloud

Registered and Segmented Deformable Object Reconstruction fr...

引用

IEEE Workshop on Applications of Computer Vision (WACV)

作者： Pit Henrich Balázs Gyenes Paul Maria Scheikl Gerhard Neumann Franziska Mathis-Ullrich Dep. Artificial Intelligence in Biomedical Engineering FAU Erlangen-Nürnberg Erlangen Germany Institute for Anthropomatics and Robotics - Karlsruhe Institute of Technology Karlsruhe Germany

In deformable object manipulation, we often want to interact with specific segments of an object that are only defined in non-deformed models of the object. We thus require a system that can recognize and locate these segments in sensor data of deformed real world objects. This is normally done using deformable object registration, which is problem specific and complex to tune. Recent methods utilize neural occupancy functions to improve deformable object registration by registering to an object reconstruction. Going one step further, we propose a system that in addition to reconstruction learns segmentation of the reconstructed object. As the resulting output already contains the information about the segments, we can skip the registration process. Tested on a variety of deformable objects in simulation and the real world, we demonstrate that our method learns to robustly find these segments. We also introduce a simple sampling algorithm to generate better training data for occupancy learning.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Using Quadcopter for Vision-based Tracking of High-Rise Vertical Structure

Using Quadcopter for Vision-based Tracking of High-Rise Vert...

引用

IEEE/ASME International Conference on Mechtronic and Embedded Systems and Applications, MESA

ISBN: (数字)9798331516239

ISBN: (纸本)9798331516246

This research paper focuses on developing visionbased autonomous navigation and motion control algorithms for quadcopters to track high-rise industrial vertical structures. A camera, fixed on the drone will capture the Red-Green-Blue (RGB) channel image and will feed to the image processing node, which is a convolution neural network (CNN) based encoder-decoder model for image segmentation. The segmented image will go to the classical computer vision algorithm to get the desired features for navigation and the motion control algorithm. In this paper, we have implemented a nonlinear sliding mode control algorithm for motion control and discuss the results of positioning and motion control of the drone for the motion in circular and spiral paths. The entire experimentation is performed on the DJI M100 drone with the Nvidia Jetson Xavier NX as an onboard computer.

关键词： Image segmentation Computer vision Spirals Tracking Navigation Inspection Motion control Sliding mode control Quadrotors Drones

来源：评论

学校读者我要写书评

暂无评论

Design and Implementation of Pipeline Detection Robot with Three-Dimensional Scanning Capability

Design and Implementation of Pipeline Detection Robot with T...

引用

IEEE International Conference on Signal Processing, Communications and Computing (ICSPCC)

作者： Dansheng Chen Min Zhao Ning Ding Xiaoqiang Yuan Nan Li Zehao Fang Center for Special Robots Shenzhen Institute of Artificial Intelligence and Robotics for Society Shenzhen China Institute of Robotics and the Intelligent Manufacturing The Chinese University of Hong Kong Shenzhen China

To adapt to different application scenarios, the structure and execution components of pipeline robots need to be customized according to specific projects. Custom design requires a lot of time and manpower. Therefore, this paper proposes a pipeline robot based on network communication, which can achieve three-dimensional inspection inside pipelines with good expandability, reducing the development work for new projects and allowing for quick product updates. The hardware of the robot is designed around an ethernet switch controller, which connects communication modules and sensors, the expansion interface is reserved for further use. The software is based on the special workflow of pipeline detection robots and is designed around robot control - data acquisition - pipeline defect analysis. Testing in various scenarios showed that the design of the robot meets the functional requirements of pipeline detection as well as having good expandability.

关键词：

来源：评论

学校读者我要写书评

暂无评论

DiffusionDrive: Truncated Diffusion Model for End-to-End Autonomous Driving

arXiv

引用

arXiv 2024年

作者： Liao, Bencheng Chen, Shaoyu Yin, Haoran Jiang, Bo Wang, Cheng Yan, Sixu Zhang, Xinbang Li, Xiangyu Zhang, Ying Zhang, Qian Wang, Xinggang Institute of Artificial Intelligence Huazhong University of Science & Technology China School of EIC Huazhong University of Science & Technology China Horizon Robotics China

Recently, the diffusion model has emerged as a powerful generative technique for robotic policy learning, capable of modeling multi-mode action distributions. Leveraging its capability for end-to-end autonomous driving is a promising direction. However, the numerous denoising steps in the robotic diffusion policy and the more dynamic, open-world nature of traffic scenes pose substantial challenges for generating diverse driving actions at a real-time speed. To address these challenges, we propose a novel truncated diffusion policy that incorporates prior multi-mode anchors and truncates the diffusion schedule, enabling the model to learn denoising from anchored Gaussian distribution to the multi-mode driving action distribution. Additionally, we design an efficient cascade diffusion decoder for enhanced interaction with conditional scene context. The proposed model, DiffusionDrive, demonstrates 10× reduction in denoising steps compared to vanilla diffusion policy, delivering superior diversity and quality in just 2 steps. On the planning-oriented NAVSIM dataset, with aligned ResNet-34 backbone, DiffusionDrive achieves 88.1 PDMS without bells and whistles, setting a new record, while running at a real-time speed of 45 FPS on an NVIDIA 4090. Qualitative results on challenging scenarios further confirm that DiffusionDrive can robustly generate diverse plausible driving actions. © 2024, CC BY-NC-SA.

关键词： Gaussian distribution

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：