检索结果-内蒙古大学图书馆

IEEE TRANSACTIONS ON INTELLIGENT VEHICLES 2024年第2期9卷 3123-3126页

作者： Yu, Hongkai Liu, Xinyu Tian, Yonglin Wang, Yutong Gou, Chao Wang, Fei-Yue Cleveland State Univ Dept Elect & Comp Engn Cleveland OH 44115 USA Cleveland State Univ Dept Comp Sci Cleveland OH 44115 USA Chinese Acad Sci Inst Automat State Key Lab Management & Control Complex Syst Beijing 100045 Peoples R China Sun Yat sen Univ Sch Intelligent Syst Engn Shenzhen Campus Shenzhen 518063 Peoples R China

There are a large number of functional sensors installed on the modern intelligent vehicles. Many Artificial Intelligence based foundation models have been proposed for smart sensing to recognize the known object classes in the new but similar scenarios. However, it is still challenging for the foundation models of smart sensing to detect all the object classes in both seen and unseen scenarios. This letter aims at pushing the boundary of smart sensing research for intelligent vehicles. We first summarize the current widely-used foundation models and the foundation intelligence needed for smart sensing of intelligent vehicles. We then explain Sora-based Parallel Vision to boost the foundation models of smart sensing from basic intelligence (1.0) to enhanced intelligence (2.0) and final generalized intelligence (3.0). Several representative case studies are discussed to show the potential usages of Sora-based Parallel Vision, followed by its future research direction.

关键词： Intelligent vehicles smart sensing sora-based parallel vision

来源：评论

学校读者我要写书评

暂无评论

Design and control of an Underactuated Finger Exoskeleton for Assisting Activities of Daily Living

引用

IEEE-ASME TRANSACTIONS ON MECHATRONICS 2022年第5期27卷 2699-2709页

作者： Li, Houcheng Cheng, Long Sun, Ning Cao, Ran Chinese Acad Sci State Key Lab Management & Control Complex Syst Inst Automat Beijing 100190 Peoples R China Univ Chinese Acad Sci Sch Artificial Intelligence Beijing 100049 Peoples R China

In this article, a novel underactuated finger exoskeleton is designed to assist grasping tasks for the elderly with weak muscle strength. In mechanical design, the human finger's phalanges and joints are considered as part of the kinematic chains to realize the human-robot kinematic compatibility. The proposed finger exoskeleton achieves the finger preshaping and grasps objects with generic shapes. The proposed exoskeleton is both actively and passively backdriveable. Moreover, the weight of the wearable part of the proposed exoskeleton is 127 g, and the overall weight is 476 g, which indicates the proposed exoskeleton is lightweight and portable. To improve the grasping performance, the multiobjective genetic algorithm is implemented to optimize contact forces, which is to maximize the sum of the forces exerted on the index finger phalanges by the exoskeleton and to minimize the difference between the contact forces. After optimization, the sum of contact forces is risen by 15%, and the difference between forces is decreased by 53%. Furthermore, the admittance control is applied to make the proposed finger exoskeleton more compliant in the preshaping phase, and the admittance control is also implemented to achieve the fingertip grasping force control in the grasping phase. Finally, experiments have been conducted to verify the range of motion, grasping forces, and feasibility of the proposed index finger exoskeleton. The effectiveness of the control algorithm has also been verified by experiments.

关键词： Admittance control assistance finger exoskeleton preshaping shape-adaptive grasping underactuation

来源：评论

学校读者我要写书评

暂无评论

A Cross-Scale and Illumination Invariance-Based Model for Robust Object Detection in Traffic Surveillance Scenarios

引用

IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION systems 2023年第7期24卷 6989-6999页

作者： Lu, Yan-Feng Gao, Jing-Wen Yu, Qian Li, Yi Lv, Yi-Sheng Qiao, Hong Chinese Acad Sci State Key Lab Multimodal Artificial Intelligence S Beijing 100190 Peoples R China 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 Nanchang Univ Sch Informat Engn Nanchang 330031 Peoples R China

Robust object detection methods in traffic surveillance scenarios often encounters challenges due to large-scale deformations and illumination variations in outdoor scenes. To enhance the tolerance of such methods against these variations, we design a cross-scale and illumination-invariant detection model (CSIM) based on the You Only Look Once (YOLO) architecture. A main cause of false detection in large-scale detection tasks is the inconsistency between various feature scales. To address this issue, we introduce an adaptive cross-scale feature fusion model to ensure the consistency of the constructed feature pyramid. To overcome the influence of uneven light, we build an illumination-invariant chromaticity space on the CSIM model, which is independent of the correlated color temperature. In addition, we adopt spatial attention modules, K-means clustering and the Mish activation function for further model optimization. The obtained experimental results show that the proposed CSIM produces excellent detection results for addressing the challenges derived from large-scale deformations and the illumination changes encountered during traffic surveillance. Compared with state-of-the-art object detection methods on public datasets, our proposed model has achieved competitive results in robust object detection tasks in traffic surveillance scenarios.

关键词： Feature extraction Object detection Lighting Traffic control Adaptation models Task analysis Robustness Traffic detection surveillance scenarios spatial feature fusion illumination invariance

来源：评论

学校读者我要写书评

暂无评论

A Multimodal Fusion Model for Estimating Human Hand Force Comparing surface electromyography and ultrasound signals

引用

IEEE ROBOTICS & AUTOMATION MAGAZINE 2022年第4期29卷 10-24页

作者： Zou, Yongxiang Cheng, Long Li, Zhengwei Univ Chinese Acad Sci Sch Artificial Intelligence Beijing 100049 Peoples R China Inst Automat Chinese Acad Sci State Key Lab Management & Control Complex Syst Beijing 100190 Peoples R China

Biomimetic robots have received significant attention in recent years. Among them, the wearable exoskeleton, which imitates the functions of the musculoskeletal system to assist humans, is a typical biomimetic robot. Given that safe human-robot interaction plays a critical role in the successful application of wearable exoskeletons, this work studies the clinical readiness of a multimodal fusion model that estimates hand force based on the surface electromyography (sEMG) and A-mode ultrasound signals of the forearm muscles. The proposed multimodal fusion model affords the biomimetic hand exoskeleton assisting the elderly in completing daily tasks or quantitatively assessing the recovery level of poststroke patients. The suggested fusion model is called Optimization of Latent Representation for the Self-Attention Convolutional Neural Network (OLR-SACNN), which utilizes a common component extraction module (CCEM) and a complementary component retention module (CCRM) to optimize latent representation of the multiple modalities. Then the optimized latent representations are fused with the self-attention mechanism. The experiments conducted on a self-collected multimodal data set verify performance of the proposed OLR-SACNN model. Specifically, compared to solely employing sEMG or A-mode ultrasound signals, the force estimation's normalized mean-square error (NMSE) based on the multiple modalities decreases by 97.7 and 38.92%, respectively. Furthermore, the OLR-SACNN model has been used to estimate the hand force of some poststroke patients and attained the desired performance.

关键词： Force Ultrasonic imaging Muscles Robots Feature extraction Estimation Exoskeletons

来源：评论

学校读者我要写书评

暂无评论

Online Minimax Q Network Learning for Two-Player Zero-Sum Markov Games

引用

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING systems 2022年第3期33卷 1228-1241页

作者： Zhu, Yuanheng Zhao, Dongbin 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

The Nash equilibrium is an important concept in game theory. It describes the least exploitability of one player from any opponents. We combine game theory, dynamic programming, and recent deep reinforcement learning (DRL) techniques to online learn the Nash equilibrium policy for two-player zero-sum Markov games (TZMGs). The problem is first formulated as a Bellman minimax equation, and generalized policy iteration (GPI) provides a double-loop iterative way to find the equilibrium. Then, neural networks are introduced to approximate Q functions for large-scale problems. An online minimax Q network learning algorithm is proposed to train the network with observations. Experience replay, dueling network, and double Q-learning are applied to improve the learning process. The contributions are twofold: 1) DRL techniques are combined with GPI to find the TZMG Nash equilibrium for the first time and 2) the convergence of the online learning algorithm with a lookup table and experience replay is proven, whose proof is not only useful for TZMGs but also instructive for single-agent Markov decision problems. Experiments on different examples validate the effectiveness of the proposed algorithm on TZMG problems.

关键词： Games Nash equilibrium Mathematical model Markov processes Convergence Dynamic programming Training Deep reinforcement learning (DRL) generalized policy iteration (GPI) Markov game (MG) Nash equilibrium Q network zero sum

来源：评论

学校读者我要写书评

暂无评论

BNAS-v2: Memory-Efficient and Performance-Collapse-Prevented Broad Neural Architecture Search

引用

IEEE TRANSACTIONS ON systems MAN CYBERNETICS-systems 2022年第10期52卷 6259-6272页

作者： Ding, Zixiang Chen, Yaran Li, Nannan Zhao, Dongbin 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

In this article, we propose BNAS-v2 to further improve the efficiency of broad neural architecture search (BNAS), which employs a broad convolutional neural network (BCNN) as the search space. In BNAS, the single-path sampling-updating strategy of an overparameterized BCNN leads to terrible unfair training issue, which restricts the efficiency improvement. To mitigate the unfair training issue, we employ a continuous relaxation strategy to optimize all paths of the overparameterized BCNN simultaneously. However, continuous relaxation leads to a performance collapse issue that leads to the unsatisfactory performance of the learned BCNN. For that, we propose the confident learning rate (CLR) and introduce the combination of partial channel connections and edge normalization. Experimental results show that 1) BNAS-v2 delivers state-of-the-art search efficiency on both CIFAR-10 (0.05 GPU days, which is 4x faster than BNAS) and ImageNet (0.19 GPU days) with better or competitive performance;2) the above two solutions are effectively alleviating the performance collapse issue;and 3) BNAS-v2 achieves powerful generalization ability on multiple transfer tasks, e.g., MNIST, FashionMNIST, NORB, and SVHN. The code is available at https://***/zixiangding/BNASv2.

关键词： Computer architecture Training Graphics processing units Optimization Convolution Topology Task analysis Broad neural architecture search (BNAS) confident learning rate (CLR) continuous relaxation image classification partial channel connections (PC)

来源：评论

学校读者我要写书评

暂无评论

Learning Accurate and Stable Point-to-Point Motions: A Dynamic System Approach

引用

IEEE ROBOTICS AND AUTOMATION LETTERS 2022年第2期7卷 1510-1517页

作者： Zhang, Yu Cheng, Long Li, Houcheng Cao, Ran 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

This letter proposes a dynamic system approach to learn point-to-point motions while keeping the stability of the dynamic system. The proposed approach is grounded on a Learning from Demonstration (LfD) method based on a neural network, which gets a better reproduction performance while guaranteeing the generalization ability. The proposed approach has been experimentally validated on the LASA dataset and by the "pick-and-place" task of Franke Emika robot, and experimental results demonstrate that: (1) compared with the state-of-the-art results, the trajectory generated by the proposed approach achieves higher accuracy (approximately 24.79%) in terms of the similarity with respect to the demonstration;(2) the proposed approach can handle high dimensional data and learn from one or more demonstrations;(3) the proposed approach can guarantee the performance regardless of the variation of starting points even in the case of high dimensional complex motions.

关键词： Point-to-point tasks neural network dynamic system generalization performance high dimensional data

来源：评论

学校读者我要写书评

暂无评论

Relentless False Data Injection Attacks Against Kalman-Filter-Based Detection in Smart Grid

引用

IEEE TRANSACTIONS ON control OF NETWORK systems 2022年第3期9卷 1238-1250页

作者： Liu, Yifa Cheng, Long 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

As one of the most dangerous cyber attacks in smart grids, the false data injection attacks pose a serious threat to power system security. To detect the false data, the traditional residual method and other improved methods, such as the Kalman-filter-based detector, have been proposed. However, these methods often have defects, especially in a very complex networked system with noises. By investigating the tolerance to the uncertainty in the residual detection method and properties of noises, the attack magnitude planning has been presented to hide the attack behind noises, which can bypass the residual detection method. As to the Kalman-filter-based detector, this article designs a specific attack strategy that can successfully deceive the Kalman-filter-based detector. Under this strategy, the false data injected at each step are used to balance the anomalies caused by previous false data, making the system look quite normal in monitoring, while deviating the system from normal operation eventually.

关键词： Attack sequence false data injection Kalman filter smart grid security state estimation

来源：评论

学校读者我要写书评

暂无评论

Data-efficient image captioning of fine art paintings via virtual-real semantic alignment training

引用

NEUROCOMPUTING 2022年第0期490卷 163-180页

作者： Lu, Yue Guo, Chao Dai, Xingyuan Wang, Fei-Yue 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

The image captioning of fine art paintings aims at generating a sentence to describe the painting content. Compared with photographic images, there are few annotated data of clear and precise content descriptions for paintings. Besides, painting images usually have abstract expressions, making it hard to extract their representative features. In this paper, we propose a virtual-real semantic alignment training process to address these challenges in painting captioning. To provide sufficient training data, we generate a virtual painting captioning dataset by applying style transfer to a large-scale photographic image captioning dataset and maintaining their annotations. To tackle the difficulty of abstract expressions, we employ a semantic alignment loss between photographic image features and virtual painting features to guide the training of the painting feature extractor. We evaluate our method in two data-hungry scenarios where only a few or no annotated painting data for training. According to the evaluation results on a public painting captioning dataset and our annotated painting captioning dataset, our model achieves significant improvements and higher data efficiency than the baselines in the two data-hungry scenarios on all datasets. (C) 2022 Elsevier B.V. All rights reserved.

关键词： Image captioning Fine art paintings Semantic alignment

来源：评论

学校读者我要写书评

暂无评论

Rethinking Appearance-Based Deep Gait Recognition: Reviews, Analysis, and Insights From Gait Recognition Evolution

引用

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING systems 2025年 PP卷 PP页

作者： Li, Jingqi Zhang, Yuzhen Zeng, Yi Ye, Changxin Xu, Wenzheng Ben, Xianye Wang, Fei-Yue Zhang, Junping Fudan Univ Sch Comp Sci Shanghai 200433 Peoples R China Shandong Univ Shenzhen Res Inst Shenzhen 518063 Peoples R China Shandong Univ Sch Informat Sci & Engn Jinan 250100 Peoples R China Obuda Univ DeSci Ctr Parallel Intelligence H-1034 Budapest Hungary Macau Univ Sci & Technol Intelligent Syst Robot & Automat Lab Macau Peoples R China CASIA State Key Lab Management & Control Complex Syst Beijing 100190 Peoples R China

Gait recognition is a prominent biometric recognition technique extensively employed in public security. Appearance-based and model-based gait recognition are two categories of methods commonly used. Specifically, appearance-based methods, which use silhouettes to represent body information, typically outperform model-based methods that rely on skeleton data, making them more popular. Recently, the shift from single-frame templates to multiframe silhouettes has advanced appearance-based gait recognition with better spatiotemporal representation. However, there is a notable lack of comprehensive studies that deepen the understanding of multiframe appearance-based gait recognition methods. This article reviews various methods to trace the evolution of gait recognition. Furthermore, we unify various performant models in one framework, study the overlooked effects on data arrangement, and explore the scaling ability of existing methods. Besides the advancement in gait recognition, we also summarize the current challenges and future prospects to foster future research.

关键词： Gait recognition Feature extraction Data models Reviews Training Skeleton Shape Analytical models Deep learning Clothing Appearance-based biometric deep learning gait recognition GaitBase

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：