检索结果-内蒙古大学图书馆

International Conference on Methods and Models in Automation and robotics (MMAR)

作者： Kamil Kasprzyk Adam Gałuszka Department of Automatic Control and Robotics Silesian University of Technology Gliwice Poland

The aim of this paper is to check whenever usage of sequence based neural networks for predicting compressed air demand can be useful in screw compressor room supervisory control systems. Industrial enterprises frequently employ compressed air systems to generate the compressed air needed for daily operations. Data was gathered from three different compressor rooms with different air demand characteristics and configuration over the period of one month. Then data was prepared, analyzed, trained and tested followed by simulation tests which determined usefulness of trained networks. Since nowadays high energy prices force energy saving build of the screw compressor itself the purpose of this text was to check if there is any room for optimization in less modern and also modern applications.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Power-of- Two Quantized YOLO Network for Pedestrian Detection with Dynamic Vision Sensor

Power-of- Two Quantized YOLO Network for Pedestrian Detectio...

引用

Euromicro Symposium on Digital System Design

作者： Dominika Przewlocka-Rus Tomasz Kryjak Department of Automatic Control and Robotics AGH University of Krakow Krakow Poland

Pedestrian detection algorithms have a wide range of applications: from video surveillance, to driver assistance systems and autonomous vehicles. The performance of these systems must be as high as possible, with minimal response time and, due to the often battery-powered operation (like in electric vehicles), low energy consumption. When designing such solutions, we therefore face challenges typical for embedded vision systems: the problem of fitting algorithms of high memory and computational complexity into small low-power devices. In this paper, we propose a system based on a Dynamic Vision Sensor (DVS), which has low power requirements and operates well in conditions with variable illumination. It is these features that may make event cameras a preferential choice over frame cameras in some applications. To ensure high accuracy, for pedestrian detection we use the YOLO (You Only Look Once) deep network on event data representation. Due to the high complexity of the applied algorithm, we propose a low precision architecture: the weights of the convolution layers are quantized logarithmically to 4 bits powers-of-two values (PoT quantization). Such compression reduces not only the memory complexity by almost 8x, but also the computational complexity by replacing most multiplication operations with bit-shifting, due to use of powers-of-two weights. At the same time, the proposed system achieves the accuracy on par with the floating point baseline, of mAP0.5 0.708.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Memory-Efficient Graph Convolutional Networks for Object Classification and Detection with Event Cameras 26

Memory-Efficient Graph Convolutional Networks for Object Cla...

引用

26th IEEE Signal Processing: Algorithms, Architectures, Arrangements, and Applications, SPA 2023

作者： Jeziorek, Kamil Pinna, Andrea Kryjak, Tomasz AGH University of Krakow Embedded Vision Systems Group Department of Automatic Control and Robotics Krakow Poland Sorbonne Universite CNRS LIP6 ParisF-75005 France

ISBN: (纸本)9798350304985

Recent advances in event camera research emphasize processing data in its original sparse form, which allows the use of its unique features such as high temporal resolution, high dynamic range, low latency, and resistance to image blur. One promising approach for analyzing event data is through graph convolutional networks (GCNs). However, current research in this domain primarily focuses on optimizing computational costs, neglecting the associated memory costs. In this paper, we consider both factors together in order to achieve satisfying results and relatively low model complexity. For this purpose, we performed a comparative analysis of different graph convolution operations, considering factors such as execution time, the number of trainable model parameters, data format requirements, and training outcomes. Our results show a 450-fold reduction in the number of parameters for the feature extraction module and a 4.5-fold reduction in the size of the data representation while maintaining a classification accuracy of 52.3%, which is 6.3% higher compared to the operation used in state-of-the-art approaches. To further evaluate performance, we implemented the object detection architecture and evaluated its performance on the N-Caltech101 dataset. The results showed an accuracy of 53.7% mAP@0.5 and reached an execution rate of 82 graphs per second. © 2023 Division of Signal Processing and Electronic Systems, Poznan University of Technology (DSPES PUT).

关键词： Cameras

来源：评论

学校读者我要写书评

暂无评论

PointPillars Backbone Type Selection for Fast and Accurate LiDAR Object Detection

PointPillars Backbone Type Selection for Fast and Accurate L...

引用

International Conference on Computer Vision and Graphics, ICCVG 2022

作者： Lis, Konrad Kryjak, Tomasz Embedded Vision Systems Group Computer Vision Laboratory Department of Automatic Control and Robotics AGH University of Science and Technology Al. Mickiewicza 30 Krakow30-059 Poland

ISBN: (纸本)9783031220241

3D object detection from LiDAR sensor data is an important topic in the context of autonomous cars and drones. In this paper, we present the results of experiments on the impact of backbone selection of a deep convolutional neural network on detection accuracy and computation speed. We chose the PointPillars network, which is characterised by a simple architecture, high speed, and modularity that allows for easy expansion. During the experiments, we paid particular attention to the change in detection efficiency (measured by the mAP metric) and the total number of multiply-addition operations needed to process one point cloud. We tested 10 different convolutional neural network architectures that are widely used in image-based detection problems. For a backbone like MobilenetV1, we obtained an almost 4x speedup at the cost of a 1.13% decrease in mAP. On the other hand, for CSPDarknet we got an acceleration of more than 1.5x at an increase in mAP of 0.33%. We have thus demonstrated that it is possible to significantly speed up a 3D object detector in LiDAR point clouds with a small decrease in detection efficiency. This result can be used when PointPillars or similar algorithms are implemented in embedded systems, including SoC FPGAs. The code is available at https://***/vision-agh/pointpillars-backbone. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Spatial modelling of virus transfer and exposure using Bayesian inference with Integrated Nested Laplace Approximation

Spatial modelling of virus transfer and exposure using Bayes...

引用

International Conference on Methods and Models in Automation and robotics (MMAR)

作者： Andrzej Neścior Adrian Dudek Waldemar Bauer Jerzy Baranowski Department of Automatic Control & Robotics AGH University of Krakow Krakow Poland

Airborne transmission is a key element in the spread of viral contagion. To prevent this, health organization publish guidelines for every major disease outbreak. However, they are often based on researches carried out without access to modern solutions. In this paper, we propose usage of Bayesian inference as an additional, to computationally-intensive methods such as CFD or FEM, way to analyse short - range virus exposure. We build a spatial model, using INLA package, which allows us to optimize the complicated computational process and deal with conundrums of virus exposure modeling.

关键词：

来源：评论

学校读者我要写书评

暂无评论

PID controller Tuning for Capsubot with Standard and Fractional-Order PSO Algorithm

PID Controller Tuning for Capsubot with Standard and Fractio...

引用

International Conference on Methods and Models in Automation and robotics (MMAR)

作者： Artur Babiarz Department of Automatic Control and Robotics Silesian University of Technology Gliwice Poland

The presented paper contains comparison two algorithms for searching optimal value of PID controllers: particle swarm optimization (PSO) and fractional-order particle swarm optimization (F-PSO) for the control system of capsubot robot. A cost function which is used with PSO and F-PSO is presented. At the end, the obtained simulation results are shown and discussed.

关键词：

来源：评论

学校读者我要写书评

暂无评论

FREQUENCY DRIVEN SCREW COMPRESSORS STATE PREDICTION IN MODERN AIR control SYSTEM 38

FREQUENCY DRIVEN SCREW COMPRESSORS STATE PREDICTION IN MODER...

引用

38th Annual European Simulation and Modelling Conference, ESM 2024

作者： Kasprzyk, Kamil Galuszka, Adam Department of Automatic Control ad Robotics Silesian University of Technology Akademicka 16 Gliwice44-100 Poland Marani Sp. z o.o Szybowa 14C Zabrze41-808 Poland

ISBN: (纸本)9789492859334

The purpose of this paper is to evaluate whether neural networks can be effectively applied in compressed air supervisory control systems to provide capabilities beyond those of conventional onsite logic controllers. The study was conducted over the course of a year in a modern compressor room system consisting of three screw compressors equipped with permanent magnet motors. During this period, production levels varied from month to month. Criteria for evaluating the accuracy of neural network predictions were proposed, along with a procedure to identify an effective model. A neural network based on Long Short-Term Memory (LSTM) architecture, featuring two LSTM layers, was implemented with a dedicated loss function. This straightforward solution yielded results with minimal computational power requirements. While not all results were suitable for decision-making, in some cases, reliable and safe solutions were obtained. Periods with fewer decision-making opportunities and more consistent compressed air demand resulted in easier and more accurate future condition predictions. © 2024 Modelling and Simulation 2024 - 38th Annual European Simulation and Modelling Conference 2024, ESM 2024. All rights reserved.

关键词： Screws

来源：评论

学校读者我要写书评

暂无评论

Two approaches (GC-ECD and electrochemical sensors signals processing) to the determination of carbonyl compounds as markers of air pollution 26

Two approaches (GC-ECD and electrochemical sensors signals p...

引用

26th IEEE Signal Processing: Algorithms, Architectures, Arrangements, and Applications, SPA 2023

作者： Bronczyk, Karolina Adamski, Michal Dabrowska, Agata Konieczka, Adam Dabrowski, Adam Adam Mickiewicz University Faculty of Chemistry Department of Trace Analysis Poznań Poland Institute of Automatic Control and Robotics Poznan University of Technology Division of Signal Processing and Electronic Systems Poznań Poland

ISBN: (纸本)9798350304985

This article presents the application possibilities of unmanned aerial vehicles (UAVs) for the rapid and efficient determination of carbonyl compounds: formaldehyde and acetaldehyde concentrations in gaseous environmental samples. One of the described methods is the use of special electrochemical sensors (PMS5003 and SEN0231) to measure air pollution in real time (every 10 seconds). The second method involves adsorbing gas samples into sorption tubes with XAD-2 cartridges and uses chromatographic analysis: gas chromatography coupled with a selective electron capture detector (GC-ECD). The proposed analytical technique allows the determination of carbonyl compounds at low concentration levels (μg/m3) using pre-derivatization of samples with the reagent 2,3,4,5,6-pentafluorobenzyl-hydroxylamine (PFBOA). Comparing the results obtained from electrochemical sensors and chromatographic analysis, we can conclude that there are discrepancies between the determined concentrations of formaldehyde in the analyzed gas samples. © 2023 Division of Signal Processing and Electronic Systems, Poznan University of Technology (DSPES PUT).

关键词： Formaldehyde

来源：评论

学校读者我要写书评

暂无评论

Event-Based Supervisor control for a Bascule Bridge 9

Event-Based Supervisor Control for a Bascule Bridge

引用

9th South-East Europe Design Automation, Computer Engineering, Computer Networks and Social Media Conference, SEEDA-CECNSM 2024

作者： Fragkoulis, Dimitrios G. Koumboulis, Fotis N. Tzamtzi, Maria P. Giannaris, Georgios L. National and Kapodistrian University of Athens Robotics Automatic Control and Cyber-Physical Systems Laboratory Department of Digital Industry Technologies School of Science Euripus Campus Euboea 34400 Greece National and Kapodistrian University of Athens Core Department Euripus Campus Euboea 34400 Greece

ISBN: (纸本)9798350342482

The automation of a bascule bridge, located in the Netherlands, is studied. The modeling of the system is worked out in the framework of Ramadge-Wonham and analyzed per automation component of the bridge. The desired performance of the bridge is expressed in the form of rules, that will be translated to five groups of sets of regular languages. Next, new groups of sets of regular languages will be introduced by the injection of appropriate uncontrollable events, in order to guarantee the controllability of the languages. Finally, an event-based supervisor scheme will be developed and the nonblockingness of the controlled automaton is proved. © 2024 IEEE.

关键词： Bascule bridges

来源：评论

学校读者我要写书评

暂无评论

Real-Time Multi-Object Tracking using YOLOv8 and SORT on a SoC FPGA

arXiv

引用

arXiv 2025年

作者： Danilowicz, Michal Kryjak, Tomasz Embedded Vision Systems Group Computer Vision Laboratory Department of Automatic Control and Robotics AGH University of Krakow Poland

Multi-object tracking (MOT) is one of the most important problems in computer vision and a key component of any vision-based perception system used in advanced autonomous mobile robotics. Therefore, its implementation on low-power and real-time embedded platforms is highly desirable. Modern MOT algorithms should be able to track objects of a given class (e.g. people or vehicles). In addition, the number of objects to be tracked is not known in advance, and they may appear and disappear at any time, as well as be obscured. For these reasons, the most popular and successful approaches have recently been based on the tracking paradigm. Therefore, the presence of a high quality object detector is essential, which in practice accounts for the vast majority of the computational and memory complexity of the whole MOT system. In this paper, we propose an FPGA (Field-Programmable Gate Array) implementation of an embedded MOT system based on a quantized YOLOv8 detector and the SORT (Simple Online Realtime Tracker) tracker. We use a modified version of the FINN framework to utilize external memory for model parameters and to support operations necessary required by YOLOv8. We discuss the evaluation of detection and tracking performance using the COCO and MOT15 datasets, where we achieve 0.21 mAP and 38.9 MOTA respectively. As the computational platform, we use an MPSoC system (Zynq UltraScale+ device from AMD/Xilinx) where the detector is deployed in reprogrammable logic and the tracking algorithm is implemented in the processor system. © 2025, CC BY.

关键词： System-on-chip

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：