ID 原文 译文
56038 仿真分析表明,算法与基于多滤波器并行工作的交互式多模型相比,跟踪精度相当,但运算量大大降低,因此适用于对性能和实时性要求较高的场合,有很好的应用前景。 Simulation analysis show that the algorithm and the interactive multiple model based on parallel filter work than tracking precision, but the computation is greatly reduced, thus suitable for requiring higher performance and real-time performance, has a good application prospect.
56039 GO法是评价具有多状态时序特性的复杂系统可靠性的有效方法,但GO法操作符众多、算法复杂并且缺乏工具软件支持,制约了GO法的工程应用。针对该问题,本文提出一种基于贝叶斯网络的GO法新算法。 GO method is to evaluate the reliability of complex systems with multiple state sequence characteristics and effective method, but the GO method operator numerous and complex algorithm and a lack of tool software support, restricted the GO method in engineering application. Aiming at this problem, this paper proposes a new algorithm of the GO method based on bayesian networks.
56040 首先,定义常用操作符到贝叶斯网络节点映射规则;然后,给出GO模型映射转换为贝叶斯网络的可编程流程; First, define common operators to the bayesian network node mapping rules;Then, we give the GO into bayesian network model mapping of programmable process;Finally, using the bayesian networks is a mature tools support, after quantitative solving mapping of the bayesian network model.
56041 最后,利用贝叶斯网络成熟工具支持,定量求解映射后的贝叶斯网络模型。 A new algorithm operator mapping rules and unified, model mapping transformation process simple and intuitive, easy to master by the engineers and application.
56042 新算法操作符映射规则统一,模型映射转换流程简单直观,便于工程人员掌握和应用。此外,除了能得出传统的定量结果,新算法使得GO法还具有故障推理和诊断能力。 In addition, in addition to the traditional quantitative results, the new algorithm makes the GO method also has the fault reasoning and diagnostic ability.
56043 为了解决局部线性嵌入(locally linear embedding,LLE)流形学习算法无法自适应确定重构区间和不能进行增量学习等问题,提出了一种自适应聚类增量LLE(clustering adaptively incremental LLE,C-LLE)目标识别算法。 In order to solve the locally linear embedding (LLE, locally linear embedding) manifold learning algorithm could not be determined adaptive reconstruction interval and incremental learning, this paper proposes a adaptive incremental clustering LLE (clustering adaptively incremental LLE, C - LLE) target recognition algorithm.
56044 该算法通过建立高维非线性样本集的局部线性结构聚类模型,对聚类后的类内样本采用线性重构,解决了LLE算法样本重构邻域无法自适应确定的问题;通过构建降维矩阵,解决了LLE算法无法单独对增量进行降维和无法利用增量对目标进行识别的问题。 The algorithm through the establishment of high-dimensional nonlinear partial linear structure of the sample set clustering model, the clustering samples within class after using linear refactoring, solved the LLE algorithm sample refactoring neighborhood could not be determined adaptive problems;By building a dimension reduction matrix, solve the dimensionality of LLE algorithm for incremental alone cannot use increment of target identification problem.
56045 实验表明,本文算法能够准确提取高维样本集的低维流形结构,具有较小的增量降维误差和良好的目标识别性能。 Experiments show that the algorithm can accurately extract high dimensional low dimensional manifold structure of sample set, with a smaller increment dimension reduction error and good performance of target recognition.
56046 机场跑道异物(foreign object debris,FOD)检测对飞行器的安全起降有着非常重要的意义,而机场跑道异物检测的一个关键环节是很好地抑制机场雷达图像的噪声,因此提出一种基于距离-时间维的移不变混合变换以抑制机场雷达图像的噪声。 Airport runway foreign bodies (foreign object debris, FOD) detection is very important significance for the safety of the aircraft take-off and landing, the airport runway foreign object detection is a key link in well inhibit the airport radar image noise, thus put forward a kind of based on distance - time d move the same mixing transform to suppress the airport radar image noise.
56047 首先,在雷达成像时进行离散傅里叶变换(discrete Fourier transform,DFT)和维纳滤波滤除距离维上的噪声。 First of all, to be the time when radar imaging discrete Fourier transform (discrete 'Fourier transform, the DFT) and wiener filtering filter out the noise on the distance d.