line graph刘逸飞
基于速度矢量场的无人机实时动态航路规划

基 于速 度 矢 量 场 的无 人 机 实 时动 态 航 路 规 划
李 春 涛 易小 芹 胡 木
(. 1 南京 航 空航 天 大学 自动 化 学 院 , 南京 , 10 6 201; 2 中 国 航 空 工业 集 团公 司 成都 飞机 设 计 研 究 所 , 都 ,1 0 1 . 成 60 4 )
性 问题 ; 动 态 实时规 划应 用 中 , 立 了环境 信 息 更新 方 法 , 在 确 实现 了对 动 态环 境 的描 述 。 过仿 真 验 证 , 明速 度 通 表 矢 量 场 法 能 够 根 据 动 态环 境 信 息及 时规 避 威 胁 到 达 目标 点 , 法具 有 良好 的 完备 性 和 实 时 性 , 用 于 局 域 动 态 算 适
摘 要 : 对 局 域 动 态环 境 中无 人 机 实时航 路 规 划 展 开 研 究 , 出 了一 种 基 于 速 度 矢量 场 的二 维 动 态 实 时航 路 规 针 提
划 方 法 。 过 建 立 不 同 空 间特 征 区域 速 度 场 模 型 , 通 实现 了速 度 场 驱 动 下 的 无人 机 航 路 规 划 。 中采 用虚 拟 目标 点 文 法 解 决 了速度 矢 量 场航 路 规 划 局 部 陷 阱问 题 ; 用探 测 步长 法 , 采 实现 了无 人 机 机 动 约 束 的 融 合 , 决 了航 路 可飞 解
Ab ta t An o —ier a—i a h p a nn s p o o e o n n e e ilv hce U AV )i wo d — sr c : n l e lt n me p t ln ig i r p s d f ru ma n d a ra e il ( nt —i
2020年第43卷总目次

Ⅲ
船用锅炉汽包水位内模滑模控制………………………………………………… 段蒙蒙ꎬ甘辉兵 ( 3. 83 )
三峡升船机变频器 IGBT 路故障诊断 ……………………… 孟令琦ꎬ高 岚ꎬ李 然ꎬ朱汉华 ( 3. 89 )
定航线下考虑 ECA 的船舶航速多目标优化模型 …………… 甘浪雄ꎬ卢天赋ꎬ郑元洲ꎬ束亚清 ( 3. 15 )
改进二阶灰色极限学习机在船舶运动预报中的应用………… 孙 珽ꎬ徐东星ꎬ苌占星ꎬ叶 进 ( 3. 20 )
Ⅱ
规则约束下基于深度强化学习的船舶避碰方法
………………………………… 周双林ꎬ杨 星ꎬ刘克中ꎬ熊 勇ꎬ吴晓烈ꎬ刘炯炯ꎬ王伟强 ( 3. 27 )
船用起重机吊索张力建模与计算机数值仿真 ………………………… 郑民民ꎬ张秀风ꎬ王任大 ( 4. 94 )
约束规划求解自动化集装箱码头轨道吊调度 ………………………… 丁 一ꎬ田 亮ꎬ林国龙 ( 4. 99 )
航海气象与环保
162 kW 柴油机排气海水脱硫性能
基于模糊 ̄粒子群算法的舰船主锅炉燃烧控制 ……… 毛世聪ꎬ汤旭晶ꎬ汪 恬ꎬ李 军ꎬ袁成清 ( 1. 88 )
多能源集成控制的船舶用微电网系统频率优化……………… 张智华ꎬ李胜永ꎬ季本山ꎬ赵 建 ( 1. 95 )
基于特征模型的疏浚过程中泥浆浓度控制系统设计………… 朱师伦ꎬ高 岚ꎬ徐合力ꎬ潘成广 ( 2. 74 )
基于卷积神经网络的航标图像同态滤波去雾 …………………………………………… 陈遵科 ( 4. 84 )
船用北斗导航系统终端定位性能的检测验证 …………………………………………… 吴晓明 ( 4. 89 )
《2024年融合因果注意力Transformer模型的股价预测研究》范文

《融合因果注意力Transformer模型的股价预测研究》篇一一、引言随着金融市场的日益复杂化,股价预测成为了投资者和金融分析师关注的焦点。
传统的股价预测方法大多基于统计模型和时间序列分析,但这些方法往往难以捕捉到市场中的非线性关系和长期依赖性。
近年来,深度学习在股价预测领域的应用逐渐受到关注,尤其是Transformer模型因其强大的特征提取能力,在处理时间序列数据上具有明显优势。
本文提出一种融合因果注意力的Transformer模型,用于股价预测研究。
二、相关工作回顾过去的股价预测研究主要依赖传统的统计和时间序列分析方法,如移动平均、ARIMA模型等。
然而,这些方法往往忽略了股票价格中存在的非线性关系和复杂的依赖性。
近年来,随着深度学习的发展,循环神经网络(RNN)及其变种在股价预测中取得了显著成果。
特别是Transformer模型,凭借其自注意力机制在处理时间序列数据时表现出强大的性能。
三、方法论本文提出的模型融合了因果注意力和Transformer模型。
首先,Transformer模型通过自注意力机制捕捉时间序列数据中的依赖关系。
其次,引入因果注意力机制,确保模型在预测时仅考虑历史信息而不涉及未来信息,这符合股票市场的实际情况。
1. 数据预处理在进行模型训练之前,需要对股票价格数据进行预处理,包括数据清洗、归一化等步骤。
此外,还需构建相应的特征集,如交易量、波动率等。
2. 融合因果注意力的Transformer模型本模型主要由编码器-解码器结构组成,编码器用于捕捉输入序列的依赖关系,解码器则根据编码器的输出进行预测。
在自注意力机制中融入因果注意力,确保模型在预测时仅考虑历史信息。
四、实验与分析本部分将详细介绍实验设置、数据集、实验结果及分析。
1. 数据集与实验设置实验采用真实的股票市场数据集,包括历史股价、交易量、波动率等特征。
将数据集划分为训练集和测试集,采用均方误差(MSE)作为损失函数,优化器选择Adam。
对极几何约束的单目相机在线标定方法

对极几何约束的单目相机在线标定方法
廖运茂;刘飞;方超;吴思齐;马运涛
【期刊名称】《北京测绘》
【年(卷),期】2024(38)2
【摘要】本文提出了一种基于点线特征检测的单目相机在线标定方法。
首先,采用点特征提取(ORB)算法、线特征提取(LSD)算法分别检测出序列图像中的点特征和线特征,利用快速最邻近搜索(FLANN)算法匹配出初始匹配点和初始匹配线;然后,基于两视图对极几何关系解算出相机的位置和姿态;最后,使用最小二乘优化方法解算相机内参和畸变参数,针对低纹理环境,如墙面、道路等人工建造物,提出了点线特征联合检测策略,提高了两相邻图像特征的匹配精度,对于只利用图像中特征点的相机标定方法,参加解算的特征点解算出来的畸变参数往往不能代表整张图像的畸变参数,为了进一步提高解算精度,对图像进行均匀切割且利用非极大值抑制来提取关键点。
实验表明,本文提出的相机在线标定方法在自然环境中有较高的精度,能够为视觉测量系统提供真实的相机参数。
【总页数】7页(P257-263)
【作者】廖运茂;刘飞;方超;吴思齐;马运涛
【作者单位】北京建筑大学测绘与城市空间信息学院;北京建筑大学科学技术发展研究院;北京建筑大学理学院
【正文语种】中文
【中图分类】P258
【相关文献】
1.一种多目相机的外参数两步联合在线标定方法
2.单目数码相机的几何标定方法
3.基于Halcon的单目相机标定方法与测量实验
4.基于改良天鹰优化器的单目相机标定方法
5.基于单目相机和多线激光雷达的联合标定方法研究
因版权原因,仅展示原文概要,查看原文内容请购买。
基于灰色神经网络模型的企业碳排放峰值预测

基 于 灰 色 神 经 网 络 模 型 的 企 业 碳 排 放 峰 值 预 测
o f c a r b o n e mi s s i o n s,wh i c h c a n no t f u l l y c o n s i d e r t h e i n f l ue n c e f a c t o r s o f c a r b o n e mi s s i o n s i n t he p r o c e s s o f p r e d i c t i o n,l e a d i n g t o t h e pr e d i c — t i o n o f t h e p e a k v a l ue o f c a r b o n e mi s s i o n s .Th e r e f o r e,t h i s p a p e r p u t s f o r wa r d a ne w mo d e l o f c a r b o n e mi s s i o n p r e d i c t i o n b a s e d o n g r e y n e u r a l
n e t wo r k mo de 1 . Th e mo d e l i s b a s e d o n t h e g r e y mo d e l ,a n d n e u r a l ne t wo r k b y c o mb i n i n g g r e y n e u r a l n e t wo r k mo d e l ,t h e c or p o r a t e c a r b o n r a w d a t a a r e s u p e r i mp o s e d,a n d i s r e p r e s e n t e d b y d i f f e r e nt i a l e q u a t i o n s ,t h e VS TE a l g o r i t h m a s t h e b a s i c a l g o r i t h m gr e y n e u r a l n e t wo r k p r e — d i c t i o n mo d e l ,t h e c a l c ul a t i o n o f c o r p o r a t e c a r b o n e mi s s i o n s p a t h o f c a r b o n e mi s s i on s ,t o me e t t h e r a n d om Ga u s s d i s t r i b u t i o n f u n c t i o n i n o r d e r
【计算机应用与软件】_算法改进_期刊发文热词逐年推荐_20140724

科研热词 遗传算法 粗糙集 小波变换 图像分割 入侵检测 阈值 边缘检测 蚁群算法 粒子群优化算法 神经网络 聚类 数据挖掘 属性约简 图像融合 关联规则 量子粒子群优化算法 透明度(α 通道) 连续属性 边界区域 算法 离散化 目标提取 旅行商问题 故障诊断 层次聚类 可靠性 动态的 优化 高速网络 高斯混合模型 高分辨率图像 骨骼动画 题库管理 频繁项集 预测模式 项集 音高间隔 音高向量 面部特征 非结构化p2p 非线性补偿 附加动量因子 阵发性心脏病 阴性选择 遗传变异算子 选择算子 选址优化 迭代阈值法 迭代算法 运行时间分布 过分割 边缘模板匹配
2008年 序号 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52
科研热词 遗传算法 蚁群算法 边缘检测 算法 文本分类 图像分割 人脸检测 车辆路径问题 自适应 聚类算法 粗糙集 粒子群优化算法 簇 特征选取 特征提取 熵 模拟退火算法 服务质量 最大-最小蚂蚁系统 旅行商问题 改进遗传算法 密度 多目标优化 入侵检测系统 dsp 高斯变异 高斯分布 页面聚类推荐算法 非线性函数 非支配集 非参数检验 零运动预判断 零树结构 降阶 降维 阈值选取 阈值分割 阈值 链接测试 遗传操作 速度的影响 逐点插入 选播 进化计算 进化算法 运动估计 边界追踪 轮廓线 轮廓形状 转换矩阵 转弯处 距离变换
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
一种Colored WF_logic Net的工作流过程建模
关键词 : 工作 流;工作流过程建模 ;着 色 P ti er 网;有 色工作流逻辑 网
W o k o Pr c s o ei gBa e n Coo e F l g cNe r f w o e sM d l s d o lr dW l n o i t
_
W ANG i g Xi LIYa M e , J n — n, n- i XU u n J a
务过程抽象 出来 , 用一种形式化的、 并 计算机可处理的
方式来表示 ,这种形式化结果就称为工作流模型…。工 作流模 型是 目前 工作流 技术研 究的热 点 ,过 程模 型是
其核心 。本 文主要探 讨工作 流过 程模 型的建立。可支 持过程 管理 的建模工具主要有 :D F系列方法 、 er IE P ti
A b ta t A f rito cn h e r l rd Per Ne n L. e ,h spa e r p s sac lr d wo k o l i sr c : t nr du i gt et o yof e h Co o e tj t d W a _ tt i p rp o o e o o e r f w n l ogc n tt c iv he m o e i fwo kfo e o a h e e t d l ng o r l w.I a e su b u a d i g i s r n e e c a ii r p s n a c s tdy a o th n ln n u a c ,r a h b l y g a h i t
ห้องสมุดไป่ตู้
( c o l fE e tr is n fr t nEn ie rn , a z o ioo gUnv ri , a z o 3 0 0 Chn ) S h o lcon c dI omai gn eig L n h uJa tn i est L n h u7 0 7 , i a o a n o y
Synopsys PrimeWave Design Environment用户指南说明书
DATASHEET OverviewSynopsys PrimeWave Design Environment is a comprehensive and flexible environment for simulation and reliability analysis of analog, RF, mixed-signal design, custom-digital and memory designs within the Synopsys Custom Design Family.It delivers a seamless simulation experience around all the engines of Synopsys PrimeSim with comprehensive analysis, improved productivity, and ease of use. PrimeWave Design Environment also offers a powerful Tcl-based scripting capability enabling easy regressions across thousands of corners. It offers a unified workflow driven analysis for all Synopsys PrimeSim Reliability Analysis applications. It provides a consistent and rich verification experience across applications. Synopsys PrimeWave Design Environment offers guided analysis setup, intuitive visualization, debug, and root-causing capabilities enabling a reliability aware design methodology.Figure 1: PrimeWave Design EnvironmentUnified WorkflowDriven Environmentfor Simulation andReliability AnalysisPrimeWave Design EnvironmentKey Benefits• Unified workflow across all PrimeSim engines for all types of simulation and reliability analysis• Tightly integrated with Custom Compiler™ for analog, RF, and mixed-signal analysis or available in a command-line mode for custom-digital and memory flows• High-capacity waveform viewing capable of handling large waveform files in a multitude of file formats.• Efficient post-processing with more than a hundred built-in functions and support for HSPICE measure statements• Open environment for integration and customization• Flexible Tcl-based scripting capability for programming complex testbenches for regression and post-processing Comprehensive Environment for Analog/RF and Mixed Signal SimulationFigure 2: PrimeWave provides a comprehensive environment for analog, RF and mixed-signal simulationThe PrimeWave Design Environment is a comprehensive environment for all Synopsys simulation engines and analysis capabilities. It provides an easy-to-use simulation setup cockpit, support for grid-based job distribution and monitoring of simulation jobs, anda powerful graphical waveform viewer. Whether designing analog, mixed-signal, or RF designs on mature or advanced nodes, the PrimeWave Design Environment is a unified solution for all applications.For mixed-signal applications, the PrimeWave Design Environment offers support for both analog and digital waveforms.For RF designs such as VCOs, LNAs, and Mixers, the PrimeWave Design Environment offers built-in phase-noise, jitter, intermodulation distortion, and gain compression measurement functions.For signal integrity, the PrimeWave Design Environment supports built-in measurements for statistical eye simulations as well as specific applications such as PAM4 and DDR. With the sequential testbench feature, the PrimeWave Design Environment allows dependencies to be created between testbenches and measurements, allowing measurements from one testbench to be used in another for a specification-driven flow. The powerful parameterization capability allows any design variable or view to be transformed into a variable that the design environment can iterate through, to allow exhaustive characterization of any design.Unified Workflow Driven Environment For Reliability AnalysisFor reliability, the PrimeWave Design Environment offers a unified workflow driven setup with several features for out-of-box and customizable workflows, advanced debug, and visualizations, distributing and monitoring compute intensive jobs.203/16/23.CS1071072274-PrimeWave-Design-Environment-DS.©2023 Synopsys, Inc. All rights reserved. Synopsys is a trademark of Synopsys, Inc. in the United States and other countries. A list of Synopsys trademarks isavailable at /copyright.html . All other names mentioned herein are trademarks or registered trademarks of their respective owners.Automated Setup WizardSchematic cross-probing & annotationConsolidated view for setup/violation/waiversDesign Schematic Design SetupWaiversViolationOA View (CCK data save/restore)Figure 3: PrimeWave Design Environment offers a unified workflow driven setup with several features for customizable workflowsSome key highlights include.• Guided wizard driven setup for new users and expert mode for advanced users.• Load, visualize and cross-probe and annotate violations in the schematic and layout open access database.• Consolidated open access view saving setup, annotations, and violations.• Advanced debugging features to quickly root-cause and isolate reliability bottlenecks.Flexible and Programmable EnvironmentA modern design environment requires a robust mechanism for running batch-mode simulations, reliability analysis and regressions. The PrimeWave Design Environment is open and extensible and can be controlled in either GUI or Batch mode with scripting. The GUI is also extensible, allowing CAD teams to craft custom measurements and create bind key settings and release them across their organizations. The PrimeWave Design Environment allows any simulation setup to easily be saved as a Tcl script, allowing all the capabilities in interactive mode to be also used in batch-mode. Users can also define pre-and post-run procedures for performing calibrations and other processing of simulation data. The environment also provides flexibility to users to customize their own checks for reliability applications. It includes a rich set of open APIs that make it easy to customize the user experience.Foundry-certified, ISO26262 Compliant and Cloud Ready• The PrimeWave Design Environment supports all PrimeSim simulators for Analog, RF and Mixed Signal simulations and offers applications for full life cycle reliability verification.• PrimeSim and PrimeSim Reliability Analysis technologies are part of the ISO 26262 TCL1 certified Synopsys Custom Design tool chain and thus can be reliably used for ASIL-D applications.• PrimeSim simulation engines and PrimeSim Reliability Analysis technologies are also cloud ready with enablement andoptimization for leading public cloud platforms.For more information about Synopsys products, support services or training, visit us on the web at , contact your local sales representative or call 650.584.5000。
《2024年融合因果注意力Transformer模型的股价预测研究》范文
《融合因果注意力Transformer模型的股价预测研究》篇一一、引言股价预测是金融市场分析和投资决策的关键环节,随着人工智能技术的飞速发展,利用深度学习模型进行股价预测成为了研究热点。
传统的股价预测方法大多依赖于线性回归、时间序列分析等统计方法,但这些方法往往无法充分捕捉股票市场中的非线性、复杂依赖关系。
近年来,注意力机制和Transformer模型在自然语言处理等领域取得了显著成果,将其应用于股价预测领域也引起了广泛关注。
本文提出了一种融合因果注意力的Transformer 模型,以更好地捕捉股价序列的时序依赖性和因果关系,提高股价预测的准确性。
二、相关文献综述在股价预测领域,已有众多研究利用不同的人工智能模型进行探索。
其中,循环神经网络(RNN)及其变体如长短期记忆网络(LSTM)和门控循环单元(GRU)被广泛应用于时间序列预测。
然而,这些模型在捕捉长期依赖关系时存在局限性。
Transformer模型因其自注意力机制在许多任务上取得了卓越的成果,近年来也被引入到股价预测中。
但现有研究鲜少考虑股价预测中的因果关系。
因此,本研究旨在融合因果注意力与Transformer模型,以提高股价预测的精度。
三、融合因果注意力Transformer模型本研究提出的模型融合了因果注意力和Transformer模型,通过自注意力和因果注意力共同作用,更好地捕捉股价序列的时序依赖性和因果关系。
模型结构包括编码器、解码器以及嵌入层等部分。
编码器利用Transformer的自注意力机制捕捉输入序列的内部关系,而因果注意力则确保模型在生成预测时考虑到了时间上的先后顺序和因果关系。
四、实证研究本研究选取了某股票市场的历史股价数据作为实验数据集,通过融合因果注意力的Transformer模型进行训练和预测。
在实验过程中,我们对比了传统统计方法、LSTM、GRU以及未融合因果注意力的Transformer模型,以验证本模型的优越性。
一种单像素攻击分布可视化方法[发明专利]
专利名称:一种单像素攻击分布可视化方法 专利类型:发明专利 发明人:孙健,王婉怡,陈杰 申请号:CN2020110664 99.8 申请日:20200930 公开号:C一种单像素攻击分布可视化方法,首先采用差分进化算法对原始图像的各像素 点逐一进行单像素攻击,得到所有能够被攻击成功的像素点,克服了现有单像素攻击每次仅仅能够收 敛至某一局部最优点而导致的随机性缺陷;同时,在全图扫描过程中,本发明采用位置攻击图M1记 录被成功攻击的像素点的位置,即单像素攻击的分布,从而能够直观且全面的展示出原始图像上可成 功攻击的像素点的位置分布和数量,实现单像素攻击分布的可视化,为用户提供攻击策略的参考信 息,对单像素攻击与神经网络可解释性的进一步研究有着非常重要的意义。
申请人:北京理工大学 地址:100081 北京市海淀区中关村南大街5号 国籍:CN 代理机构:北京理工大学专利中心 更多信息请下载全文后查看