检索结果-内蒙古大学图书馆

Vehicle Detection and Tracking Based on Improved YOLOv8

IEEE ACCESS 2025年 13卷 24793-24803页

作者： Liu, Yunxiang Shen, Shujun Shanghai Inst Technol Dept Comp Sci & Informat Engn Shanghai 201418 Peoples R China

With the rapid increase in transportation pressure, the demand for efficient traffic recognition and tracking systems is growing. Traditional methods have certain limitations when dealing with complex situations in traffic scenes, such as large weights and insufficient detection accuracies, et al. Therefore, we propose a novel method based on YOLOv8n. Firstly we introduced SCC_Detect based on SCConv on the detection head to reduce the computation of redundant features. Then we replaced the convolutional kernel with a dual convolutional kernel to construct a lightweight deep neural network. Subsequently, the Focaler-EIoU loss function is introduced to improve the accuracy. The BotSORT tracker is embedded in the period of inference, which achieves more accurate and stable recognition and tracking results in the traffic scene. The experimental results show that the proposed model reduces parameters and weight by approximately 36.5% and 25% respectively at the expense of only 0.2% mAP@0.5 compared with YOLOv8n on the UA-DETRAC dataset. In terms of tracking, MOTA, IDF1 and MOTP of the BoTSORT algorithm on the test video were superior to those of DeepSORT and ByteTrack. The accuracy was improved, and the number of lost tracks was reduced. It has a high practical value and application prospects in traffic detection and deployment.

关键词： Feature extraction Accuracy Convolution Roads Target tracking Kernel Convolutional neural networks YOLO Radar tracking inference algorithms Vehicle tracking multi-object track lightweight model YOLOv8 vehicle detection BotSORT

来源：评论

学校读者我要写书评

暂无评论

Optimization of Waiting Time for Electric Vehicles Using a Fuzzy inference System

引用

IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS 2022年第9期23卷 15396-15407页

作者： Hussain, Shahid Kim, Yun-Su Thakur, Subhasis Breslin, John G. Natl Univ Ireland Galway NUIG Data Sci Inst DSI Dept Elect & Elect Engn Galway H91 TK33 Ireland Gwangju Inst Sci & Technol GIST Grad Sch Energy Convergence Gwangju 61005 South Korea

Electric vehicles (EVs) need to be recharged at intermediate locations, such as shopping malls, restaurants, and parking lots, to meet the daily commute requirements of their users. Currently, public electric vehicle supply equipment (EVSE) serve EVs by conventional methods, which can result in long waiting time for users. This issue reduces the travel efficiency of EVs and thus affects user comfort. Most previous research has studied energy consumption and charging cost optimization;however, comparatively less work has focused on waiting time optimization despite its great importance from the EV user's perspective. In this paper, we formulate the waiting time optimization as a fuzzy integer linear programming problem and propose a novel heuristic fuzzy inference system-based algorithm (FISA) that resolves the objective function and minimizes the waiting time of EVs at public EVSE installations. We developed the underlying fuzzy inference system by defining the membership functions, expert rules, and formulation for obtaining the optimal solution. The novel FISA automates the correlations of the uncertain and independent input parameters into weighted control variables and resolves the objective function in each sampling period to optimize the waiting time for EVs with the most urgent service requirements. A java language-based simulator is developed for a parking lot to evaluate the effectiveness of the proposed FISA. The simulation results indicate higher efficiency of the proposed FISA compared with state-of-art scheduling algorithms.

关键词： Optimization Fuzzy logic Indexes Linear programming inference algorithms Costs Batteries Charging and waiting times electric vehicles electric vehicle supply equipment fuzzy integer linear programming fuzzy inference system

来源：评论

学校读者我要写书评

暂无评论

PW-YOLO-Pose: A Novel Algorithm for Pose Estimation of Power Workers

引用

IEEE ACCESS 2024年 12卷 116841-116860页

作者： Su, Qin Zhang, Juling Chen, Mingju Peng, Hongming Sichuan Univ Sci & Engn Sch Automat & Informat Engn Yibin 644000 Peoples R China Sichuan Univ Sci & Engn Artificial Intelligence Key Lab Sichuan Prov Yibin 644000 Peoples R China Power Internet Things Key Lab Sichuan Prov Chengdu 610095 Peoples R China

To address the detection challenges of keypoints, such as misdetections and omissions causedby backgrounds, occlusions, small targets, and extreme viewpoints in complex electrical power operationenvironments for power workers. This study proposes a 2D pose estimation algorithm for power workersbased on YOLOv5s6-Pose: PW-YOLO-Pose. In this study, the detection rate of occluded keypoints isimproved by embedding the Swin Transformer encoder in the top layer of the backbone network. Theproposed BiFPN (a weighted bi-directional feature pyramid network) structure with a small target detectionlayer improves the detection rate of small target characters and the precision of their keypoints'*** keypoint regression precision is improved overall by using CA (coordinate attention) in the model neckand improving the bounding box regression loss function to Wise-IoU. The algorithmic model in this studydemonstrates excellent detection and largely meets the real-time requirements on the proposed power workerpose estimation dataset in this study. ThemAP(0.5)(The mean average precision when the threshold for objectkeypoint similarity is set to 0.5.) andmAP(0.5:0.95)are 93.35% and 64.75% respectively, which are 5.22% and1.53% higher than the baseline model. The detection time of a single image is 21.3 ms, respectively. It canserve as a valuable theoretical foundation and reference for behavior recognition and state monitoring ofpower workers in intricate electrical power operation environments.

关键词： Pose estimation Prediction algorithms YOLO Feature extraction inference algorithms Detection algorithms Transformers Bidirectional control Power system reliability Personnel Encoding Electrical power operation pose estimation YOLO-Pose detection of keypoints

来源：评论

学校读者我要写书评

暂无评论

Root Cause Analysis for Cloud-Native Applications

引用

IEEE TRANSACTIONS ON CLOUD COMPUTING 2024年第1期12卷 232-250页

作者： Zurkowski, Bartosz Zielinski, Krzysztof AGH Univ Sci & Technol Fac Comp Sci PL-30059 Krakow Poland

Root cause analysis (RCA) is a critical component in maintaining the reliability and performance of modern cloud applications. However, due to the inherent complexity of cloud environments, traditional RCA techniques become insufficient in supporting system administrators in daily incident response routines. This article presents an RCA solution specifically designed for cloud applications, capable of pinpointing failure root causes and recreating complete fault trajectories from the root cause to the effect. The novelty of our approach lies in approximating causal symptom dependencies by synergizing several symptom correlation methods that assess symptoms in terms of structural, semantic, and temporal aspects. The solution integrates statistical methods with system structure and behavior mining, offering a more comprehensive analysis than existing techniques. Based on these concepts, in this work, we provide definitions and construction algorithms for RCA model structures used in the inference, propose a symptom correlation framework encompassing essential elements of symptom data analysis, and provide a detailed description of the elaborated root cause identification process. Functional evaluation on a live microservice-based system demonstrates the effectiveness of our approach in identifying root causes of complex failures across multiple cloud layers.

关键词： Cloud computing Correlation Semantics Behavioral sciences Trajectory Measurement inference algorithms Root cause analysis event correlation knowledge mining cloud-native applications

来源：评论

学校读者我要写书评

暂无评论

Block-Wise Mixed-Precision Quantization: Enabling High Efficiency for Practical ReRAM-Based DNN Accelerators

引用

IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS 2024年第12期43卷 4558-4571页

作者： Wu, Xueying Hanson, Edward Wang, Nansu Zheng, Qilin Yang, Xiaoxuan Yang, Huanrui Li, Shiyu Cheng, Feng Pande, Partha Pratim Doppa, Janardhan Rao Chakrabarty, Krishnendu Li, Hai Duke Univ Dept Elect & Comp Engn Durham NC 27708 USA Univ Virginia Dept Elect & Comp Engn Charlottesville VA 22904 USA Univ Calif Berkeley Dept Elect Engn & Comp Sci Berkeley CA 94720 USA Washington State Univ Sch Elect Engn & Comp Sci Pullman WA 99163 USA Arizona State Univ Sch Elect Comp & Energy Engn Tempe AZ 85281 USA

Resistive random access memory (ReRAM)-based processing-in-memory (PIM) architectures have demonstrated great potential to accelerate the deep neural network (DNN) training/inference. However, the computational accuracy of analog PIM is compromised due to the nonidealities, such as the conductance variation of ReRAM cells. The impact of these nonidealities worsens as the number of concurrently activated wordlines (WLs) and bitlines (BLs) increases. To guarantee computational accuracy, only a limited number of WLs and BLs of the crossbar array can be turned on concurrently, significantly reducing the achievable parallelism of the *** the constraints on parallelism limit the efficiency of the accelerators, they also provide a new opportunity for the fine-grained mixed-precision quantization. To enable efficient DNN inference on the practical ReRAM-based accelerators, we propose an algorithm-architecture co-design framework called block-wise mixed-precision quantization (BWQ). At the algorithm level, the BWQ algorithm (BWQ-A) introduces a mixed-precision quantization scheme at the block level, which achieves a high weight and activation compression ratio with negligible accuracy degradation. We also present the hardware architecture design BWQ-H, which leverages the low-bit-width models achieved by BWQ-A to perform high-efficiency DNN inference on the ReRAM devices. BWQ-H also adopts a novel precision-aware weight mapping method to increase the ReRAM crossbar's throughput. Our evaluation demonstrates the effectiveness of BWQ, which achieves a 6.08x speedup and a 17.47x energy saving on average compared to the existing ReRAM-based architectures.

关键词： Quantization (signal) Computer architecture Computational modeling Artificial neural networks Performance evaluation inference algorithms Training Deep neural network (DNN) acceleration mixed-precision quantization model compression processing-in-memory (PIM) resistive random access memory (ReRAM)

来源：评论

学校读者我要写书评

暂无评论

Explain the World-Using Causality to Facilitate Better Rules for Fuzzy Systems

引用

IEEE TRANSACTIONS ON FUZZY SYSTEMS 2024年第12期32卷 6671-6683页

作者： Zhang, Te Wagner, Christian Garibaldi, Jonathan M. Univ Nottingham Sch Comp Sci Lab Uncertainty Data & Decis Making LUCID & Intell Nottingham NG8 1BB England Univ Nottingham Sch Comp Sci Intelligent Modelling & Anal Res Grp Nottingham NG8 1BB England

The rules of a rule-based system provide explanations for its behavior by revealing the relationships between the variables captured. However, ideally, we have AI systems which go beyond explainable AI (XAI), that is, systems which not only explain their behavior, but also communicate their "insights" with respect to the real world. This requires rules to capture causal relationships between variables. In this article, we argue that those systems where the rules reflect causal relationships between variables represent an important class of fuzzy rule-based systems with unique benefits. Specifically, such systems benefit from improved performance and robustness;facilitate global explainability and thus cater to a core ambition for AI: the ability to communicate important relationships among a system's real-world variables to the human users of AI. We establish two causal-rule focused approaches to design fuzzy systems, and show the distinctions in their respective application scenarios for the explanations of the rules obtained by these two methods. The results show that rules which reflect causal relationships are more suitable for XAI than rules which "only" reflect correlations, while also confirming that they offer robustness to over-fitting, in turn supporting strong performance.

关键词： Fuzzy systems Artificial intelligence Cause effect analysis inference algorithms Complexity theory Semantics Data models Causal explanations causality rule-based systems XAI

来源：评论

学校读者我要写书评

暂无评论

YOLOV9-CBM: An Improved Fire Detection Algorithm Based on YOLOV9

引用

IEEE ACCESS 2025年 13卷 19612-19623页

作者： Geng, Xin Han, Xiao Cao, Xianghong Su, Yixuan Shu, Dongxue Zhengzhou Univ Light Ind Coll Bldg Environm Engn Zhengzhou 450001 Henan Peoples R China Shanghai Univ Sch Mechatron Engn & Automat Shanghai 200444 Peoples R China

Regarding the current problems of false alarms and missed detections in fire detection, we propose a high-precision fire detection algorithm, YOLOV9-CBM (C3-SE, BiFPN, MPDIoU), by optimizing YOLOV9. Firstly, to tackle the shortage of both quality and quantity in the existing fire datasets, we collected 2,000 fire and smoke images to establish a dataset named CBM-Fire. Secondly, the RepNCSPELAN4 module of the YOLOv9 backbone was replaced with the C3 module containing SE Attention to improve detection efficiency while guaranteeing accuracy. Besides, we transformed the multi-scale fusion network PANet in the baseline algorithm into a bidirectional feature network pyramid BiFPN to facilitate the bidirectional flow of features, enabling the algorithm to fuse information at different scales more effectively. Finally, instead of CIoU losses, we adopted MPDIoU losses in bounding box regression, which improved the accuracy of model regression and classification. Experimental results indicate that compared with YOLOV9, the recall rate of YOLOV9-CBM has increased by 7.6% and the mAP has risen by 3.8%. The revised model demonstrates good generalization performance and robustness. Code and dataset are at https://***/GengHan-123/***.

关键词： Accuracy Feature extraction Attention mechanisms YOLO inference algorithms Detection algorithms Deep learning Biological system modeling Forestry Computer architecture BiFPN C3 fire detection MPDIoU SE YOLOV9

来源：评论

学校读者我要写书评

暂无评论

Real-Time Task Scheduling With Fairness in Digital Twin Systems

引用

IEEE INTERNET OF THINGS JOURNAL 2025年第7期12卷 7846-7862页

作者： Kim, Cheonyong Saad, Walid Han, Jonghun Yu, Tao Sakaguchi, Kei Jung, Minchae Korea Natl Univ Transportat Dept Comp Software Chungju 27469 South Korea Virginia Tech Dept Elect & Comp Engn Arlington VA 22203 USA Korea Railrd Res Inst Smart Elect & Signaling Div Train Control & Commun Res Team Uiwang 16105 South Korea Inst Sci Tokyo Dept Elect & Elect Engn Tokyo 1528550 Japan Sejong Univ Dept Elect & Informat Engn Seoul 05006 South Korea

Digital twin (DT) can help create a digital representation of a physical system, thereby reflecting its real-time status. The digital object, often called cyber twin (CT), facilitates real-time monitoring and control of the physical object, i.e., the so-called physical twin (PT). Owing to this ability, CTs can optimize the PTs and simulate their status, without interrupting the physical world. Given the various CT use cases, one can identify two distinct types of DT tasks: 1) update tasks for PT-CT synchronization and 2) inference tasks for obtaining real-time testing responses. The diverse real-time requirements for update/inference tasks raise the task scheduling problem that has been neglected in previous studies. In this article, the real-time DT task scheduling problem is investigated. In particular, a new approach for evaluating the performance of real-time scheduling of DT tasks is introduced considering the relationship between update/inference tasks and fairness among CTs. Moreover, offline and online DT task scheduling schemes are proposed with the goals of maximizing the DT freshness ratio and minimizing task rejections. In particular, the DT freshness ratio maximization problem is formulated as an offline task scheduling scheme. The proposed offline solution can significantly reduce the solution space without losing optimality. Furthermore, the scheduling policies for achieving the maximal DT freshness ratio are established using which an online scheduling algorithm is designed. Simulation results show that the proposed offline/online schemes increase the DT freshness ratio by at least 16% and 11%, respectively, compared to benchmarks. The results also show that the task rejection ratio of the proposed online algorithm is within 8% of the lower bound.

关键词： Real-time systems Job shop scheduling Testing Optimization Internet of Things Digital twins Synchronization inference algorithms Computational modeling Accuracy Digital twin (DT) fairness offline and online algorithm real-time task scheduling

来源：评论

学校读者我要写书评

暂无评论

STAMP: Differentiable Task and Motion Planning via Stein Variational Gradient Descent

引用

IEEE ROBOTICS AND AUTOMATION LETTERS 2025年第6期10卷 6007-6014页

作者： Lee, Yewon Li, Andrew Z. Huang, Philip Heiden, Eric Jatavallabhula, Krishna Murthy Damken, Fabian Smith, Kevin Nowrouzezahrai, Derek Ramos, Fabio Shkurti, Florian Univ Toronto Toronto ON M5S 1A1 Canada Carnegie Mellon Univ 15213USA Pittsburgh PA USA NVIDIA Santa Clara CA 95051 USA MIT Cambridge MA 02139 USA Tech Univ Darmstadt D-64289 Darmstadt Germany McGill Univ Montreal PQ H3A0G4 Canada Univ Sydney Sydney NSW 2050 Australia

Planning for sequential robotics tasks often requires integrated symbolic and geometric reasoning. TAMP algorithms typically solve these problems by performing a tree search over high-level task sequences while checking for kinematic and dynamic feasibility. This can be inefficient because, typically, candidate task plans resulting from the tree search ignore geometric information. This often leads to motion planning failures that require expensive backtracking steps to find alternative task plans. We propose a novel approach to TAMP called Stein Task and Motion Planning (STAMP) that relaxes the hybrid optimization problem into a continuous domain. This allows us to leverage gradients from differentiable physics simulation to fully optimize discrete and continuous plan parameters for TAMP. In particular, we solve the optimization problem using a gradient-based variational inference algorithm called Stein Variational Gradient Descent. This allows us to find a distribution of solutions within a single optimization run. Furthermore, we use an off-the-shelf differentiable physics simulator that is parallelized on the GPU to run parallelized inference over diverse plan parameters. We demonstrate our method on a variety of problems and show that it can find multiple diverse plans in a single optimization run while also being significantly faster than existing approaches.

关键词： Planning Physics Optimization inference algorithms Search problems Probabilistic logic Graphics processing units Trajectory Stochastic processes Mathematical models Task and motion planning probabilistic inference

来源：评论

学校读者我要写书评

暂无评论

ENCODE: Breaking the Trade-Off Between Performance and Efficiency in Long-Term User Behavior Modeling

引用

IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING 2025年第1期37卷 265-277页

作者： Zhou, Wen-Ji Zheng, Yuhang Feng, Yinfu Ye, Yunan Xiao, Rong Chen, Long Yang, Xiaosong Xiao, Jun Alibaba Grp Hangzhou 310000 Peoples R China Hong Kong Univ Sci & Technol Dept Comp Sci & Engn Kowloon Hong Kong 999077 Peoples R China Bournemouth Univ Bournemouth BH12 5BB England Zhejiang Univ Hangzhou 310027 Peoples R China

Long-term user behavior sequences are a goldmine for businesses to explore users' interests to improve Click-Through Rate (CTR). However, it is very challenging to accurately capture users' long-term interests from their long-term behavior sequences and give quick responses from the online serving systems. To meet such requirements, existing methods "inadvertently" destroy two basic requirements in long-term sequence modeling: R1) make full use of the entire sequence to keep the information as much as possible;R2) extract information from the most relevant behaviors to keep high relevance between learned interests and current target items. The performance of online serving systems is significantly affected by incomplete and inaccurate user interest information obtained by existing methods. To this end, we propose an efficient two-stage long-term sequence modeling approach, named as EfficieNt Clustering based twO-stage interest moDEling (ENCODE), consisting of offline extraction stage and online inference stage. It not only meets the aforementioned two basic requirements but also achieves a desirable balance between online service efficiency and precision. Specifically, in the offline extraction stage, ENCODE clusters the entire behavior sequence and extracts accurate interests. To reduce the overhead of the clustering process, we design a metric learning-based dimension reduction algorithm that preserves the relative pairwise distances of behaviors in the new feature space. While in the online inference stage, ENCODE takes the off-the-shelf user interests to predict the associations with target items. Besides, to further ensure the relevance between user interests and target items, we adopt the same relevance metric throughout the whole pipeline of ENCODE. The extensive experiment and comparison with SOTA on both industrial and public datasets have demonstrated the effectiveness and efficiency of our proposed ENCODE.

关键词： Predictive models Measurement Feature extraction Prediction algorithms inference algorithms Electronic commerce Clustering algorithms Knowledge engineering Data models Data mining CTR prediction user behavior modeling metric learning recommendation systems

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

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

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：