检索结果-内蒙古大学图书馆

Vision-Based Underwater Target Following control of an Agile Robotic Manta With Flexible Pectoral Fins

IEEE ROBOTICS AND AUTOMATION LETTERS 2023年第4期8卷 2277-2284页

作者： Meng, Yan Wu, Zhengxing Li, Yating Chen, Di Tan, Min Yu, Junzhi Peking Univ Coll Engn Dept Adv Mfg & Robot State Key Lab Turbulence & Complex Syst Beijing 100871 Peoples R China Univ Chinese Acad Sci Sch Artificial Intelligence Beijing 100049 Peoples R China Chinese Acad Sci Inst Automat State Key Lab Management & Control Complex Syst Beijing 100190 Peoples R China Guangdong Ocean Univ Coll Elect & Informat Engn Zhanjiang 524088 Peoples R China

Manta-inspired robots have been increasingly applied in vision-based underwater tasks. However, owing to the refraction effect and motion coupling problem, their accurate target following control is still poorly understood. This letter proposes a vision-based underwater target following scheme for robotic manta. First, an agile robotic manta equipped with a binocular camera system is constructed. The motion modes are further detailed based on the propulsion analysis to the flexible pectoral fins. Second, a color-based underwater target detection and localization method is conceived. Specifically, the binocular refraction correction is employed to improve the localization accuracy. Third, combing the nonsingular terminal sliding mode controller and fuzzy inference system, a robust target following controller is presented to solve the coupling between distance and azimuth control of the robot. Finally, various aquatic experiments are conducted, verifying the effectiveness and robustness of the proposed underwater target following scheme. The obtained results provide valuable insights into the intelligent control of agile bioinspired robots performing vision-based tasks in dynamic underwater environments.

关键词： Robot kinematics Location awareness Cameras Task analysis Robot vision systems Turning Sports Biologically-inspired robots robotic manta target following control binocular underwater localization sliding mode control

来源：评论

学校读者我要写书评

暂无评论

A Wire-driven Elastic Robotic Fish and its Design and CPG-Based control

引用

JOURNAL OF INTELLIGENT & ROBOTIC systemS 2023年第1期107卷 1-13页

作者： Liao, Xiaocun Zhou, Chao Wang, Jian Fan, Junfeng Zhang, Zhuoliang Chinese Acad Sci Inst Automat State Key Lab Management & Control Complex Syst 95 Zhongguancun East Rd Beijing 100190 Peoples R China Univ Chinese Acad Sci Sch Artificial Intelligence 19 Yuquan Rd Beijing 100049 Peoples R China

To simulate the complex and continuous undulation of fishtail in nature, the method of adopting the discrete Multi-Joint mechanism requires a certain number of degrees of freedom, which results in the complexities of mechanism and control necessarily. Compared with Multi-Joint, flexible tail is a better scheme due to continuum, robustness, and simpler control. Hence, this paper proposes a wire-driven elastic robotic fish with flexible tail, which simulates fish muscle through multi-wire drive and adopts a fishlike spine design based on elastic component. Due to these distinctive designs, our robotic fish not only realizes the compliant simulation of fishlike swimming gait, but also owns higher bionic degree. Further, the kinematics model and speed estimation model of the wire-driven elastic robotic fish are developed, and the error between the body wave and the desired fishlike swimming gait is further optimized so as to determine the appropriate parameters of central pattern generator. The results show that the optimized body wave of fishtail matches well with the desired fishlike swimming gait and the RMSE of the stable swimming speed between simulations and experiments is 0.045 m/s, which validates the proposed model and optimization method. Finally, the relationships between the frequency and swimming speed under small amplitude are explored, from which we find that amplitude has a greater impact on speed than frequency at high frequency, and the maximum swimming speed of about 0.54 m/s, i.e., 1.02 BL/s, is obtained.

关键词： Robotic fish Wire-driven mode Elastic component Kinematics model Body wave

来源：评论

学校读者我要写书评

暂无评论

Invariant Extended Kalman Filtering for Tightly Coupled LiDAR-Inertial Odometry and Mapping

引用

IEEE-ASME TRANSACTIONS ON MECHATRONICS 2023年第4期28卷 2213-2224页

作者： Shi, Pengcheng Zhu, Zhikai Sun, Shiying Zhao, Xiaoguang Tan, Min Chinese Acad Sci Inst Automat State Key Lab Management & Control Complex Syst Beijing 100190 Peoples R China Univ Chinese Acad Sci Sch Artificial Intelligence Beijing 100049 Peoples R China NIO Inc Shanghai 201804 Peoples R China

In this article, we extend the invariant extended Kalman filter (EKF) to light detection and ranging (LiDAR)-inertial odometry and mapping systems using invariant observer design and the theory of Lie groups for directly fusing LiDAR and inertial measurement unit (IMU) measurements. We consider this from two different aspects and implement two independent systems. Specifically, we propose a robo-centric invariant EKF LiDAR-inertial odometry termed Inv-LIO1. Its mapping module is an ordinary used one and two modules run in separate threads. A world-centric invariant EKF LiDAR-inertial odometry termed Inv-LIO2 is designed and implemented, which has an integrated odometry and mapping architecture. In Inv-LIO1, the output of the filter is the pose estimated by the scan-to-scan match method, which serves as the initial estimate of the mapping module that refines the odometry and constructs a 3-D map. The robo-centric formulation represents that the state in a local frame shifted at every LiDAR time to prevent filter divergence. Inv-LIO2 directly fuses LiDAR feature points and IMU data to obtain the map-refined odometry by scan-to-map match method, followed by global map update. To validate the effectiveness and robustness of the proposed method, we conduct extensive experiments in various indoor and outdoor environments. Overall, Inv-LIO1 offers pure odometry with higher accuracy than other state-of-the-art systems, improving the overall performance. Inv-LIO2 achieves superior accuracy over other state-of-the-art systems in the map-refined odometry comparison.

关键词： Laser radar Feature extraction Simultaneous localization and mapping Robustness Point cloud compression Optimization Kalman filters Invariant extended kalman filter (EKF) light detection and ranging (LiDAR)-inertial odometry multisensor fusion localization state estimation

来源：评论

学校读者我要写书评

暂无评论

Towards Better Generalization of Deep Neural Networks via Non-Typicality Sampling Scheme

引用

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING systemS 2023年第10期34卷 7910-7920页

作者： Peng, Xinyu Wang, Fei-Yue Li, Li Tsinghua Univ Dept Automat Beijing 100084 Peoples R China Chinese Acad Sci State Key Lab Management & Control Complex Syst Inst Automat Beijing 100080 Peoples R China Tsinghua Univ Tsinghua Natl Lab Informat Sci & Technol Dept Automat Beijing 100084 Peoples R China

Improving the generalization performance of deep neural networks (DNNs) trained by minibatch stochastic gradient descent (SGD) has raised lots of concerns from deep learning practitioners. The standard simple random sampling (SRS) scheme used in minibatch SGD treats all training samples equally in gradient estimation. In this article, we study a new data selection method based on the intrinsic property of the training set to help DNNs have better generalization performance. Our theoretical analysis suggests that this new sampling scheme, called the nontypicality sampling scheme, boosts the generalization performance of DNNs through biasing the solution toward wider minima, under certain assumptions. We confirm our findings experimentally and show that more variants of minibatch SGD can also benefit from the new sampling scheme. Finally, we discuss an extension of the nontypicality sampling scheme that holds promise to enhance both generalization performance and convergence speed of minibatch SGD.

关键词： Training Estimation Deep learning Standards Optimization Noise measurement Convergence Deep learning generalization performance nontypicality sampling scheme stochastic gradient descent (SGD)

来源：评论

学校读者我要写书评

暂无评论

Short-term wind power forecast with turning weather based on DBSCAN-RFE-LightGBM

引用

RENEWABLE ENERGY 2025年 251卷

作者： Wu, Zhenlong Fan, Xinyu Bian, Guibin Liu, Yanhong Zhang, Xiaoke Chen, Yangquan Zhengzhou Univ Sch Elect & Informat Engn Zhengzhou 450001 Peoples R China Zhengzhou Univ State Key Lab Intelligent Agr Power Equipment Luoyang 471000 Peoples R China Chinese Acad Sci Inst Automat State Key Lab Management & Control Complex Syst Beijing 100190 Peoples R China State Grid Henan Elect Power Co Elect Power Res Inst Zhengzhou 450052 Peoples R China Univ Calif Merced Mechatron Embedded Syst & Automation MESA Lab Dept Mech Engn Sch Engn Merced CA 95343 USA

At present, the wind power prediction technology is becoming more and more mature, however, the turning weather brings a decrease in prediction accuracy due to its suddenness and instability. In this paper, a wind power prediction method considering the turning period and non-turning period respectively is proposed to weaken the adverse effects caused by turning weather. Firstly, the prediction effects of multiple models are compared, density-based spatial clustering of applications with noise (DBSCAN) is selected as the outlier processing method, recursive feature elimination (RFE) is used as the feature selection method, and Light gradient boosting machine (LightGBM) is used for prediction. The combined prediction method based on DBSCAN-RFE-LightGBM can reduce the influence of abnormal data and redundant features and improve the prediction effect. Then, the sliding window is set to detect the turning period. Considering that the turning weather is an emergency and does not happen frequently, the amount of data is small, which leads to the inability to train the model well. Generative adversarial networks (GAN) are applied to expand the turning period data. The LightGBM is trained and predicted by using the expended turning period data. Finally, the time-division prediction results are merged. Using the data collected from wind farms for short-term power prediction experiments, the time-segment prediction method proposed in this paper with GAN reduces MAE by 1.913 and RMSE by 3.351 on a single unit compared with the non-differentiated period.

关键词： Turning weather Combination model Generative adversarial networks Multi-unit Wind power forecast DBSCAN-RFE-lightGBM

来源：评论

学校读者我要写书评

暂无评论

Depth-Aware Multi-Person 3D Pose Estimation With Multi-Scale Waterfall Representations

引用

IEEE TRANSACTIONS ON MULTIMEDIA 2023年 25卷 1439-1451页

作者： Shen, Tianyu Li, Deqi Wang, Fei-Yue Huang, Hua Beijing Normal Univ Sch Artificial Intelligence Beijing 100875 Peoples R China Chinese Acad Sci State Key Lab Management & Control Complex Syst Beijing 100190 Peoples R China Chinese Acad Sci Inst Automat Beijing 100190 Peoples R China

Estimating absolute 3D poses of multiple people from monocular image is challenging due to the presence of occlusions and the scale variation among different persons. Among the existing methods, the top-down paradigms are highly dependent on human detection which is prone to the influence from inter-person occlusions, while the bottom-up paradigms suffer from the difficulties in keypoint feature extraction caused by scale variation and unreliable joint grouping caused by occlusions. To address these challenges, we introduce a novel multi-person 3D pose estimation framework, aided by multi-scale feature representations and human depth perceiving. Firstly, a waterfall-based architecture is incorporated for multi-scale feature representations to achieve a more accurate estimation of occluded joints with a better detection of human shapes. Then the global and local representations are fused for handling the effects of inter-person occlusion and scale variation in depth perceiving and keypoint feature extraction. Finally, with the guidance of the fused multi-scale representations, a depth-aware model is exploited for better 2D joint grouping and 3D pose recovering. Quantitative and qualitative evaluations on benchmark datasets of MuCo-3DHP and MuPoTS-3D prove the effectiveness of our proposed method. Furthermore, we produce an occluded MuPoTS-3D dataset and the experiments on it validate the superiority of our method for overcoming the occlusions.

关键词： Three-dimensional displays Pose estimation Feature extraction Location awareness Cameras Semantics Solid modeling Human depth perceiving multi-person 3d pose estimation multi-scale representation occlusion handling

来源：评论

学校读者我要写书评

暂无评论

A light defect detection algorithm of power insulators from aerial images for power inspection

引用

NEURAL COMPUTING & APPLICATIONS 2022年第20期34卷 17951-17961页

作者： Yang, Lei Fan, Junfeng Song, Shouan Liu, Yanhong Zhengzhou Univ Sch Elect Engn Zhengzhou 450001 Peoples R China Chinese Acad Sci Inst Automat State Key Lab Management & Control Complex Syst Beijing 100190 Peoples R China

With the rapid growth of high-voltage transmission lines, the number of power transmission line equipments is correspondingly increasing. Power insulator is the basic component which plays the key role in the stable operation of power system. As a common defect of power insulators, missing-cap issue will affect the structural strength and durability of different power insulators. Therefore, the condition monitoring of power insulators is a daily but priority power line inspection task. Faced with the weak image features of small insulator defects in the aerial images, the conventional handcrafted features could not extract effectively powerful image features. Meanwhile, the small-scale insulator defects will bring a certain effect to the model training of deep learning. Therefore, the high-efficiency and accurate defect inspection still present a challenging task against complex backgrounds. To address the above issues, aimed at the missing-cap defects of power insulators, a novel defect identification algorithm from aerial images is proposed by taking advantage of state-of-the-art deep learning and transfer learning models. Fused with Spatial Pyramid Pooling (SPP) and MobileNet networks, a light deep convolutional neural network (DCNN) model based on You Only Look Once (YOLO) V3 network is proposed for fast and accurate insulator location to remove complex background interference. On the basis, combined with Dempster-Shafer (DS) evidence theory, the improved transfer learning model based on feature fusion is proposed for high-precision defect identification of power insulators. Experiments show that the proposed method could acquire a better identification performance against complex power inspection environment compared with other related detection models.

关键词： Insulator location Defect identification Transfer learning Dempster-Shafer evidence theory

来源：评论

学校读者我要写书评

暂无评论

Cross-modality synergy network for referring expression comprehension and segmentation

引用

NEUROCOMPUTING 2022年 467卷 99-114页

作者： Li, Qianzhong Zhang, Yujia Sun, Shiying Wu, Jinting Zhao, Xiaoguang Tan, Min Chinese Acad Sci Inst Automat State Key Lab Management & Control Complex Syst Beijing Peoples R China Univ Chinese Acad Sci Sch Artificial Intelligences Beijing Peoples R China

Referring expression comprehension and segmentation aim to locate and segment a referred instance in an image according to a natural language expression. However, existing methods tend to ignore the interaction between visual and language modalities for visual feature learning, and establishing a synergy between the visual and language modalities remains a considerable challenge. To tackle the above problems, we propose a novel end-to-end framework, Cross-Modality Synergy Network (CMS-Net), to address the two tasks jointly. In this work, we propose an attention-aware representation learning module to learn modal representations for both images and expressions. A language self-attention submodule is proposed in this module to learn expression representations by leveraging the intra-modality relations, and a language-guided channel-spatial attention submodule is introduced to obtain the language aware visual representations under language guidance, which helps the model pay more attention to the referent-relevant regions in the images and relieve background interference. Then, we design a cross-modality synergy module to establish the inter-modality relations for modality fusion. Specifically, a language-visual similarity is obtained at each position of the visual feature map, and the synergy is achieved between the two modalities in both semantic and spatial dimensions. Furthermore, we propose a multi-scale feature fusion module with a selective strategy to aggregate the important information from multi-scale features, yielding target results. We conduct extensive experiments on four challenging benchmarks, and our framework achieves significant performance gains over state-of-the-art methods. (c) 2021 Elsevier B.V. All rights reserved.

关键词： Referring expression comprehension Referring expression segmentation Cross-modality synergy Attention mechanism

来源：评论

学校读者我要写书评

暂无评论

A speed measurement method for underwater robots using an artificial lateral line sensor

引用

SMART MATERIALS AND STRUCTURES 2022年第1期31卷 015011-015011页

作者： Zhang, Zhuoliang Zhou, Chao Cao, Zhiqiang Tan, Min Cheng, Long Deng, Sai Fan, Junfeng Chinese Acad Sci Inst Automat State Key Lab Management & Control Complex Syst Beijing Peoples R China Univ Chinese Acad Sci Sch Artificial Intelligence Beijing Peoples R China

Underwater robot technology has made considerable progress in recent years. However, due to the harsh environment and noise in the flow field near the underwater robots, it is difficult to measure some basic parameters, including swimming speed. The traditional speed measurement methods for underwater robots have the disadvantages of being limited by the environment and bulky. In order to overcome these shortcomings, an artificial lateral line (ALL) sensor based on cantilever structure was developed in this paper. According to the deformation of cantilever beam under water impact, the swimming speed of underwater robots can be measured. In addition, an 'end-to-end' calibration algorithm was proposed to calibrate the ALL sensor in the noisy environment, avoiding the complicated noise modeling and filter design process. To reduce the risk of overfitting, a hybrid loss function based on physical model was adopted. Compared with the classical calibration method, our method can reduce the error by 47.8%. Our sensor achieved an average absolute error of 0.07897 m s(-1), and can measure water speed up to 3 m s(-1).

关键词： artificial lateral line speed of underwater robot hybrid loss function calibration in noisy environment flow sensor

来源：评论

学校读者我要写书评

暂无评论

Advances in Flexible Robotic Manipulator systems-Part I: Overview and Dynamics Modeling Methods

引用

IEEE-ASME TRANSACTIONS ON MECHATRONICS 2024年第2期29卷 1100-1110页

作者： Li, Bai Li, Xinyuan Gao, Hejia Wang, Fei-Yue Hunan Univ State Key Lab Adv Design & Mfg Vehicle Body Changsha 410082 Peoples R China Hunan Univ Coll Mech & Vehicle Engn Changsha 410082 Peoples R China Anhui Univ Sch Artificial Intelligence Hefei 230601 Peoples R China Chinese Acad Sci State Key Lab Management & Control Complex Syst Beijing 100190 Peoples R China Chinese Acad Sci Inst Automat Beijing 100190 Peoples R China

Flexible robotic manipulators (FRMs) exhibit advantages over traditional rigid-body manipulators, including lower energy consumption, faster response, enhanced safety, reduced spatial footprint, and greater operational flexibility. This article reviews the advancements in the modeling, planning, and control technologies of FRMs, along with their applications and perspectives. Compared with previous survey papers about FRMs, this work has the following highlights. First, modeling methods w.r.t. joint flexibility, link flexibility, and link-joint coupling effects are systematically reviewed;second, FRM motion planning methods are reviewed solely instead of being treated as part of FRM control methods;third, this survey presents the recent advancements in FRMs that incorporate artificial intelligence (AI), acknowledging AI as a prevalent research trend;and finally, this is the only survey to systematically explore the rapidly evolving applications of FRM systems over the past five years. Our work is divided into two independent articles. This article, which is the first one, includes the basic background of FRMs, an overview of previous FRM-related surveys, and dynamic modeling methods for FRMs. The second article will review the prevalent motion planning and control methods for an FRM, together with FRM applications and perspectives.

关键词： Artificial intelligence (AI) dynamic modeling flexibility flexible robotic manipulator (FRM) vibration suppression

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：