检索结果-南通市图书馆

Swarm intention identification via dynamic distribution probability image

在线全文

学校读者我要写书评

暂无评论

Chinese Journal of Aeronautics 2024年第10期37卷 380-392页

作者： Yinhan WANG Jiang WANG shaoming he Fei WANG Qi WANG School of Aerospace Engineering Beijing Institute of TechnologyBeijing 100081China Institute of Computing Technology Chinese Academy of SciencesBeijing 100190China

In the realm of decision-making for defense and security applications,it is paramount to swiftly and accurately identify the intentions of incoming *** identification methods predominantly focus on single-target applications and overlook the perturbations introduced by measurement *** this study,we propose a novel concept:the Dynamic Distribution Probability(DDP)image,which is constructed using the estimated state and its covariance *** grayscale pixel value within the image signifies the probability of the presence of the agent within the *** proposed identification scheme integrates the use of Extended Kalman Filter(EKF),Convolutional Neural Network(CNN),Back Propagation(BP)network,and Gated Recurrent Unit(GRU)***,the DDP image is processed through a CNN to distill the formation characteristics,and the estimated swarm state from EKF is inputted into a BP network to deduce the kinematic *** outputs from both networks are summed and subsequently channeled into a GRU network to capture temporal *** numerical simulations and flight experiments are presented to demonstrate the robust anti-noise capability of the proposed scheme compared with conventional methods,as well as its superior training efficiency.

关键词： Intention Identification Kalman filters Convolutional neural networks Swarm intelligence

A unified formulation of optimal guidance-to-collision law for accelerating and decelerating targets

在线全文

学校读者我要写书评

暂无评论

Chinese Journal of Aeronautics 2022年第7期35卷 40-54页

作者： Young-Won KIM Boseok KIM Chang-Hun LEE shaoming he Department of Aerospace Engineering Korea Advanced Institute of Science and TechnologyDaejeon 34141Republic of Korea School of Aerospace Engineering Beijing Institute of TechnologyBeijing 100081China

This paper provides a unified formulation of optimal guidance-to-collision law for a target with an arbitrary acceleration or *** collision course for general target acceleration or deceleration is first determined from the engagement geometry in conjunction with the nonlinear engagement kinematics in the proposed *** heading error defined in the collision course is then adopted as a variable to be nullified for accomplishing the intercept *** proposed guidance law is derived based on the heading error dynamics and the optimal error dynamics to ensure optimality and finite-time *** illustrative examples,the proposed guidance command for a constant target acceleration and a target deceleration in the form of a quadratic function of speed are ***,the time-to-go prediction method is suggested for implementing the proposed *** characteristics of the proposed guidance command are analytically investigated to provide insight into the proposed *** key benefits of the proposed method lie in not producing unnecessary guidance commands near a target compared to other methods and ensuring optimality in guidance command even in the nonlinear engagement ***,numerical simulations are performed to validate the proposed method and to show our findings.

关键词： Guidance-to-collision law Optimal guidance Optimal error dynamics Target acceleration or deceleration Unified formulation

Reservoir Regulation for Control of an Ancient Landslide Reactivated by Water Level Fluctuations in heishui River,China

在线全文

学校读者我要写书评

暂无评论

Journal of Earth Science 2020年第6期31卷 1058-1067页

作者： Jian Guo Mo Xu Qiang Zhang Xianxuan Xiao Shishu Zhang shaoming he State Key Laboratory of Geohazard Prevention and Geoenvironment Protection Chengdu University of TechnologyChengdu 610059China PowerChina Chengdu Engineering Corporation Limited Chengdu 610072China

Due to the complex geological processes of Qinghai-Tibet Plateau,numerous deposits,especially the large-scale ancient landslide deposits,are characteristic features of the valleys incised in southwestern *** water level fluctuations since 2011 in Maoergai Reservoir,China,registered the reactivation of Xierguazi ancient landslide,and presented a significant risk to neighboring *** on detailed field survey and drilling exploration,the landslide was divided into Zone A and Zone B,and other characterizations of landslide were studied as *** precisely measure the extent of landslide displacement during filling and drawdown stage,surface displacement monitoring system was deployed on the *** monitoring analyses data reveal that reservoir fluctuation is the dominant factor influencing landslide displacement,especially during drawdown ***,a future sliding is anticipated in Zone A,while a creep had already existed in Zone B.A reservoir regulation was then established using the lead-lag correlation between reservoir fluctuation and landslide displacement and landslide stability *** the end,the follow-up deformation monitoring demonstrates that the reservoir regulation controlled the landslide *** control by reservoir regulation in Maoergai can serve as a case study for other settlements involved in similar construction activities.

关键词： ancient landslide reactivation reservoir regulation displacement monitoring correlation analysis

Trajectory Optimization for Target Localization and Sensor Bias Calibration with Bearing-Only Information

在线全文

学校读者我要写书评

暂无评论

Guidance, Navigation and Control 2022年第3期2卷 11-29页

作者： Xiwen Yang shaoming he Hyo-Sang Shin Antonios Tsourdos School of Aerospace Engineering Beijing Institute of Technology Beijing 100081P.R.China Yangtze River Delta Graduate School of Beijing Institute of Technology Jiaxing 314019P.R.China School of Aerospace Transport and ManufacturingCranfield UniversityCranfield MK430ALUK

The problem of trajectory optimization of an unmanned aerial vehicle(UAV)for static target localization with biased bearing measurements is *** angular bias in sensor measurements is modeled as an additive constant in the observation model and jointly estimated with the position of the *** necessary conditions for system observability of this estimation problem is first derived analytically with geometrical interpretations *** trajectory of UAV is designed based on the Fisher Information Matrix(FIM)considering physical constraints to enhance the system *** results with Monte-Carlo runs are presented to demonstrate the improvement in target localization with biased measurements by UAV trajectory optimization.

关键词： Bearing-only measurement target localization bias estimation trajectory optimization

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

浅谈地铁车站空调负荷特性

暖通空调 2007年第8期37卷 125-127,120页

作者：何绍明广州市地铁设计研究院

介绍了地铁车站空调负荷的计算方法。以广州典型地铁车站的设计数据为例,经分析得出了地铁车站空调负荷的主要控制因素是车站客流量和新风负荷的结论。

关键词：地铁空调负荷屏蔽门逐时负荷

在线全文

学校读者我要写书评

暂无评论

某餐饮娱乐楼空调方案比选

暖通空调 2005年第9期35卷 93-95页

作者：何绍明广州地铁设计研究院

介绍了建筑物的概况和对空调系统的使用要求,比较分析了三种空调方案的初投资及运行费,得到了适合该建筑的空调方案。

关键词：空调方案螺杆式冷水机组 VRV空调系统空气源热泵机组比选娱乐餐饮空调系统建筑物运行费

在线全文

学校读者我要写书评

暂无评论

一种基于迁移学习的多任务制导算法

兵工学报 2024年第6期45卷 1787-1798页

作者：罗皓文何绍溟亢有为北京理工大学宇航学院北京100081 北京理工大学无人机自主控制技术北京市重点实验室北京100081 上海机电工程研究所上海201109

针对典型的飞行器制导任务,利用深度学习算法可以有效地拟合导弹飞行状态与制导指令之间的函数关系。然而当制导任务发生变化时,其二者之间的映射关系也会随之改变,从而导致在当前环境下预训练好的模型无法直接作用于新环境,重新训练制... 详细信息

针对典型的飞行器制导任务,利用深度学习算法可以有效地拟合导弹飞行状态与制导指令之间的函数关系。然而当制导任务发生变化时,其二者之间的映射关系也会随之改变,从而导致在当前环境下预训练好的模型无法直接作用于新环境,重新训练制导模型需要大量的弹道数据和巨额的时间成本。为解决上述问题,基于迁移学习的思想引入域对抗神经网络,提出基于迁移学习的多任务制导算法。以1个含有大量标签数据的源域任务辅助2个含有极少量标签数据的目标域任务进行迁移学习,从而克服预训练与在线控制之间的环境差异。使用特征提取器和域判别器提取出对任务环境不敏感的关键特征,使神经网络学习到各个任务所共享的底层信息;为提高预测精度,分别设计针对不同任务的偏置加速度预测器。数值仿真结果表明:基于迁移学习的多任务制导算法实现了导弹在不同任务中的加速度指令预测。

关键词：多约束制导计算制导深度学习迁移学习偏置比例导引

维普期刊数据库博看期刊

在线全文

学校读者我要写书评

暂无评论

基于预测校正的落角约束计算制导方法

航空学报 2022年第8期43卷 515-530页

作者：刘子超王江何绍溟李宇飞北京理工大学宇航学院北京100081 北京理工大学无人机自主控制技术北京市重点实验室北京100081 北京理工大学信息与电子学院北京100081

针对带有落角约束的导弹制导问题,提出一种基于预测校正思想的计算制导方法。根据导弹飞行状态与落角之间的映射关系设计了深度监督学习网络,在导弹飞行过程中基于观测信息实时预测落角;根据落角误差与控制指令之间的映射关系设计了深... 详细信息

针对带有落角约束的导弹制导问题,提出一种基于预测校正思想的计算制导方法。根据导弹飞行状态与落角之间的映射关系设计了深度监督学习网络,在导弹飞行过程中基于观测信息实时预测落角;根据落角误差与控制指令之间的映射关系设计了深度强化学习模型,在比例导引律的基础上增加偏置项,基于预测落角误差实时输出偏置项,校正落角误差。引入深度监督学习网络解决了传统强化学习方法中的稀疏奖励问题。设计了仿真飞行实验训练深度监督学习网络与深度强化学习模型,然后测试了提出的计算制导方法的性能。仿真结果表明所设计的计算制导方法准确地实现了落角控制,在嵌入式计算机的计算精度高,实时性好,具备工程实用价值。

关键词：落角约束计算制导预测校正制导偏置比例导引律深度学习

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

基于迁移学习的角度约束时间最短制导算法

航空学报 2023年第19期44卷 237-251页

作者：罗皓文何绍溟金天宇刘子超北京理工大学宇航学院北京100081 北京理工大学无人机自主控制技术北京市重点实验室北京100081

由于导弹所处的气动环境等外界条件复杂多变,因此在特定气动环境中表现优异的深度监督学习制导算法将无法直接作用于新环境,这给制导指令的精准预测带来了很大挑战。针对上述问题,本文以落角约束下优化飞行时间的制导问题为背景,提出了... 详细信息

由于导弹所处的气动环境等外界条件复杂多变,因此在特定气动环境中表现优异的深度监督学习制导算法将无法直接作用于新环境,这给制导指令的精准预测带来了很大挑战。针对上述问题,本文以落角约束下优化飞行时间的制导问题为背景,提出了一种基于迁移学习的角度约束时间最短制导算法,该算法只需极少量的新数据,便能使制导律迅速适应不同的气动环境。在本文中,我们首先通过域对抗地训练特征提取器和域判别器,从不同气动环境的弹道数据中提取出对气动变化不敏感的关键特征;其次设计了适应于不同气动环境的偏置加速度预测器,进而实现了对导弹的精准制导;最后,大量数值仿真结果表明,在新气动环境下,基于迁移学习的角度约束时间最短制导算法仍能实现制导指令的精准预测。

关键词：计算制导深度学习迁移学习落角约束偏置比例导引律