检索结果-内蒙古大学图书馆

NeRF in robotics: A Survey

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Wang, Guangming Pan, Lei Peng, Songyou Liu, Shaohui Xu, Chenfeng Miao, Yanzi Zhan, Wei Tomizuka, Masayoshi Pollefeys, Marc Wang, Hesheng The University of Cambridge ETH Zurich Switzerland The Department of Automation Shanghai Jiao Tong University Shanghai200240 China The Key Laboratory of System Control and Information Processing Ministry of Education of China China Engineering Research Center of Intelligent Control for Underground Space Ministry of Education School of Information and Control Engineering Advanced Robotics Research Center China University of Mining and Technology Xuzhou221116 China ETH Zurich Zurich Switzerland Mechanical Systems Control Laboratory University of California Berkeley United States

Meticulous 3D environment representations have been a longstanding goal in computer vision and robotics fields. The recent emergence of neural implicit representations has introduced radical innovation to this field as implicit representations enable numerous capabilities. Among these, the Neural Radiance Field (NeRF) has sparked a trend because of the huge representational advantages, such as simplified mathematical models, compact environment storage, and continuous scene representations. Apart from computer vision, NeRF has also shown tremendous potential in the field of robotics. Thus, we create this survey to provide a comprehensive understanding of NeRF in the field of robotics. By exploring the advantages and limitations of NeRF, as well as its current applications and future potential, we hope to shed light on this promising area of research. Our survey is divided into two main sections: The Application of NeRF in robotics and The Advance of NeRF in robotics, from the perspective of how NeRF enters the field of robotics. In the first section, we introduce and analyze some works that have been or could be used in the field of robotics from the perception and interaction perspectives. In the second section, we show some works related to improving NeRF's own properties, which are essential for deploying NeRF in the field of robotics. In the discussion section of the review, we summarize the existing challenges and provide some valuable future research directions for reference. © 2024, CC BY-NC-ND.

关键词： robotics

DATA: Domain-Aware and Task-Aware Pre-training

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Chang, Qing Peng, Junran Xie, Lingxi Sun, Jiajun Yin, Haoran Tian, Qi Zhang, Zhaoxiang University of Chinese Academy of Sciences China Huawei Inc China Institute of Automation Chinese Academy of Sciences China National Laboratory of Pattern Recognition China Center for Research on Intelligent Perception and Computing Centre for Artificial Intelligence and Robotics HKISI CAS China

The paradigm of training models on massive data without label through self-supervised learning (SSL) and fine-tuning on many downstream tasks has become a trend recently. However, due to the high training costs and the unconsciousness of downstream usages, most self-supervised learning methods lack the capability to correspond to the diversities of downstream scenarios, as there are various data domains, different vision tasks and latency constraints on models. Neural architecture search (NAS) is one universally acknowledged fashion to conquer the issues above, but applying NAS on SSL seems impossible as there is no label or metric provided for judging model selection. In this paper, we present DATA, a simple yet effective NAS approach specialized for SSL that provides Domain-Aware and Task-Aware pre-training. Specifically, we (i) train a supernet which could be deemed as a set of millions of networks covering a wide range of model scales without any label, (ii) propose a flexible searching mechanism compatible with SSL that enables finding networks of different computation costs, for various downstream vision tasks and data domains without explicit metric provided. Instantiated With MoCo v2, our method achieves promising results across a wide range of computation costs on downstream tasks, including image classification, object detection and semantic segmentation. DATA is orthogonal to most existing SSL methods and endows them the ability of customization on downstream needs. Extensive experiments on other SSL methods demonstrate the generalizability of the proposed method. Code is released at https://***/GAIA-vision/GAIA-ssl. Copyright © 2022, The Authors. All rights reserved.

关键词： Supervised learning

Pseudo-LiDAR for Visual Odometry

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Deng, Huiying Wang, Guangming Feng, Zhiheng Jiang, Chaokang Wu, Xinrui Miao, Yanzi Wang, Hesheng Advanced Robotics Research Center AI Research Institute School of Information and Control Engineering China University of Mining and Technology Xuzhou221116 China Department of Automation Key Laboratory of System Control and Information Processing of Ministry of Education Key Laboratory of Marine Intelligent Equipment and System of Ministry of Education Shanghai Engineering Research Center of Intelligent Control and Management Shanghai Jiao Tong University Shanghai200240 China

As one of the important tasks in the field of robotics and machine vision, LiDAR/visual odometry provides tremendous help for various applications such as navigation, location, etc. In the existing methods, LiDAR odometry shows superior performance, but visual odometry is still widely used for its price advantage. Conventionally, the task of visual odometry mainly rely on the input of continuous images. However, it is very complicated for the odometry network to learn the epipolar geometry information provided by the images. In this paper, the concept of pseudo-LiDAR is introduced into the odometry to solve this problem. The pseudo-LiDAR point cloud back-projects the depth map generated by the image into the 3D point cloud, which changes the way of image representation. Compared with the stereo images, the pseudo-LiDAR point cloud generated by the stereo matching network can get the explicit 3D coordinates. Since the 6 Degrees of Freedom (DoF) pose transformation occurs in 3D space, the 3D structure information provided by the pseudo-LiDAR point cloud is more direct than the image. Compared with sparse LiDAR, the pseudo-LiDAR has a denser point cloud. In order to make full use of the rich point cloud information provided by the pseudo-LiDAR, a projection-aware dense odometry pipeline is adopted. Most previous LiDAR-based algorithms sampled 8192 points from the point cloud as input to the odometry network. The projection-aware dense odometry pipeline takes all the pseudo-LiDAR point clouds generated from the images except for the error points as the input to the network. While making full use of the 3D geometric information in the images, the semantic information in the images is also used in the odometry task. The fusion of 2D-3D is achieved in an image-only based odometry. Experiments on the KITTI dataset prove the effectiveness of our method. To the best of our knowledge, this is the first visual odometry method using pseudo-LiDAR. Copyright © 2022, The Authors. A

关键词： Pipelines

Finite-Time and Fixed-Time Bipartite Containment Control of Cooperative-Antagonistic Networks: A unified Framework

学校读者我要写书评

暂无评论

TechRxiv

TechRxiv 2022年

作者： Zhu, Zhen-Hua Wu, Huaiyu Guan, Zhi-Hong Chen, Yang Zheng, Xiujuan Engineering Research Center for Metallurgical Automation and Measurement Technology of Ministry of Education Wuhan University of Science and Technology Wuhan430081 China Institute of Robotics and Intelligent Systems Wuhan University of Science and Technology Wuhan430081 China School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China

This article studies the problems of finite-time and fixed-time bipartite containment control for cooperative-antagonistic networks with multiple leaders over arbitrary weakly connected signed digraphs, in which leaders can be stationary or evolving dynamically through interacting with other leaders in their neighborhood. To this end, a unified distributed nonlinear control scheme is proposed using the nearest neighbor rule. It is shown that under the proposed control scheme, all followers can be guaranteed to converge towards the convex hull formed by each leader’s trajectory and its symmetrical one within a finite time or fixed time so long as the underlying weakly connected signed digraph has at least one structurally balanced closed strong component. The efficacy of the proposed control scheme is illustrated via simulations. © 2022, CC BY.

关键词： Directed graphs

Robust prescribed-time coordination control of cooperative-antagonistic networks with disturbances

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Zhu, Zhen-Hua Wu, Huaiyu Guan, Zhi-Hong Liu, Zhi-Wei Chen, Yang Zheng, Xiujuan The Engineering Research Center for Metallurgical Automation and Measurement Technology of Ministry of Education Wuhan University of Science and Technology Wuhan430081 China The Institute of Robotics and Intelligent Systems Wuhan University of Science and Technology Wuhan430081 China The School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China

This article targets at addressing the robust prescribed-time coordination control (PTCC) problems for single-integrator cooperative-antagonistic networks (CANs) with external disturbances under arbitrary fixed signed digraphs without any structural constraints. Toward this end, the PTCC problems for nominal single-integrator CANs without disturbances are first investigated and a fully distributed control protocol with a time-varying gain, which grows to infinity as the time approaches the settling time, is proposed utilizing the relative states of neighboring agents. Then, based on the proposed control protocol for the nominal single-integrator CANs, a new second-order prescribed-time sliding mode control protocol is constructed to achieve accurate PTCC for single-integrator CANs in the presence of external disturbances. Using Lyapunov based analysis, sufficient conditions to guarantee the prescribed-time stability, bipartite consensus, interval bipartite consensus, and bipartite containment of single-integrator CANs without or with disturbances are, respectively, derived. In the end, numerical simulations are given to confirm the derived results. © 2022, CC BY.

关键词： Sliding mode control

research on fault feature extraction of rotating machinery based on adaptive VMD and optimized CYCBD

学校读者我要写书评

暂无评论

Research on fault feature extraction of rotating machinery b...

第34届中国控制与决策会议

作者： Ran Xie Ling Xiong Binbin Dan Zeyu Ren Engineering Research Center for Metallurgical Automation and Measurement Technology of Ministry of Education Institute of Robotics and Intelligent Systems Wuhan University of Science and Technology Key Laboratory of Metallurgical Equipment and Control of Ministry of Education School of Physics and Electronic Science Hunan Institute of Science and Technology

To solve the problem that early fault features of rotating machinery are difficult to extract,an adaptive k-value hierarchical variational mode decomposition(H-VMD) combined with optimized maximum second-order cyclostationarity blind deconvolution(CYCBD) fault feature extraction method is proposed in this *** decomposition of the vibration signal is performed with *** then the noise dominant component is denoised by wavelet threshold denoising(WTD).Furthermore,the improved autocorrelation-weighted correlated kurtosis(ACK)when CYCBD enhances the periodic shock component of the denoised signal is the fitness function of ChOA,and the envelope demodulation analysis of the feature-enhanced signal is performed using the teager energy operator(TEO).Simulation analysis and experimental results show that the interference of background noise can be effectively removed and the periodic shock component of the vibration signal be enhanced by the proposed method,which is a new feature extraction method for the fault diagnosis of rotating machinery.

关键词： Variational mode decomposition Maximum second-order cyclostationarity blind deconvolution Wavelet threshold denoising Teager energy operator Fault feature extraction

Riemannian Online Optimistic Algorithms with Dynamic Regret

学校读者我要写书评

暂无评论

IEEE Transactions on Automatic Control 2025年

作者： Wang, Xi Yuan, Deming Hong, Yiguang Hu, Zihao Wang, Lei Shi, Guodong University of New South Wales School of Electrical Engineering & Telecommunications SydneyNSW2052 Australia Nanjing University of Science and Technology School of Automation Nanjing210094 China Tongji University Shanghai Research Institute for Intelligent Autonomous Systems Shanghai201210 China Hong Kong University of Science and Technology Department of Mathematics Hong Kong Zhejiang University College of Control Science and Engineering Zhejiang 310058 China The University of Sydney Australian Center for Robotics School of Aerospace Mechanical and Mechatronic Engineering NSW2006 Australia

In this paper, we consider Riemannian online convex optimization with dynamic regret, which involves minimizing the cumulative loss difference between a learner's decisions and a sequence of adaptive decisions across dynamic environments over time on Riemannian manifolds. First, we propose Riemannian Online Optimistic Gradient Descent (R-OOGD), which extends the classical optimistic algorithms to Riemannian manifolds, requiring one gradient inquiry per round. We then propose Riemannian Adaptive Online Optimistic Gradient Descent (R-AOOGD) by incorporating the meta-expert framework into R-OOGD. We analyze the dynamic regret of R-OOGD and R-AOOGD in terms of the regularity of the sequence of cost functions and comparators. The dynamic regret bound matches the results in Euclidean space and with a scale adjustment of O(1+κ D2/√1-KD2) related to the sectional curvature lower bound κ, sectional curvature upper bound and feasible set diameter. Finally, we provide numerical experiments to illustrate the performance of the proposed algorithms. © 1963-2012 IEEE.

关键词： Optimization algorithms

Event-triggered state and fault estimation of flight control system for a quadrotor UAV

学校读者我要写书评

暂无评论

Event-triggered state and fault estimation of flight control...

第33届中国控制与决策会议

作者： Xiujuan Zheng Huaiyu Wu Shiqian Zhang Engineering Research Center for Metallurgical Automation and Measurement Technology of Ministry of Education Wuhan University of Science and Technology Institute of Robotics and Intelligent Systems Wuhan University of Science and Technology

In this paper,the state and fault estimation problem is considered for a flight control system of unmanned aerial vehicle(UAV) under event-triggered transmission ***,considering the minimum variance estimation theory,an event-triggered state and fault estimator are *** the upper bounds of the estimation error covariances are obtained in the form of Rickett-type difference *** the estimator gains are obtained in realtime by minimizing the upper bounds through a recursive ***,the effectiveness and practicability of the method are verified by the flight control system of a quadrotor UAV.

关键词： UAV Event-triggered mechanism Fault estimation Minimum variance estimation

FFPA-Net: Efficient Feature Fusion with Projection Awareness for 3D Object Detection

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Jiang, Chaokang Wang, Guangming Wu, Jinxing Miao, Yanzi Wang, Hesheng Engineering Research Center of Intelligent Control for Underground Space Ministry of Education School of Information and Control Engineering Advanced Robotics Research Center China University of Mining and Technology Xuzhou221116 China Department of Automation Key Laboratory of System Control and Information Processing of Ministry of Education Key Laboratory of Marine Intelligent Equipment and System of Ministry of Education Shanghai Engineering Research Center of Intelligent Control and Management Shanghai Jiao Tong University Shanghai200240 China The Department of Engineering Mechanics Shanghai Jiao Tong University Shanghai200240 China

Promising complementarity exists between the texture features of color images and the geometric information of LiDAR point clouds. However, there still present many challenges for efficient and robust feature fusion in the field of 3D object detection. In this paper, first, unstructured 3D point clouds are filled in the 2D plane and 3D point cloud features are extracted faster using projection-aware convolution layers. Further, the corresponding indexes between different sensor signals are established in advance in the data preprocessing, which enables faster cross-modal feature fusion. To address LiDAR points and image pixels misalignment problems, two new plug- and-play fusion modules, LiCamFuse and BiLiCamFuse, are proposed. In LiCamFuse, soft query weights with perceiving the Euclidean distance of bimodal features are proposed. In BiLiCamFuse, the fusion module with dual attention is proposed to deeply correlate the geometric and textural features of the scene. The quantitative results on the KITTI dataset demonstrate that the proposed method achieves better feature-level fusion. In addition, the proposed network shows a shorter running time compared to existing methods. Copyright © 2022, The Authors. All rights reserved.

关键词： Object detection