ID 原文 译文
38646 显然,准确估计出GTD散射中心参数对刻画目标散射特性犹为重要。 Clearly and accurately estimate the GTD scattering center is important parameter to depict the target scattering characteristics.
38647 针对经典多重信号分类(multiple signal classification,MUSIC)法仅利用目标原始回波数据、参数估计精度不高这一问题,提出一种改进的MUSIC算法对散射参数估计提取。 For classical multiple signal classification (multiple signal classification, MUSIC) method USES only the raw echo data, parameter estimation precision is not high this problem, an improved MUSIC algorithm to extract scattering parameter estimation.
38648 改进的MUSIC算法通过对原始回波数据取共轭,构建新的总协方差矩阵,有效利用了目标原始回波数据的共轭信息。 Improved MUSIC algorithm based on the raw echo data of conjugate, build a new covariance matrix, the effective use of target's raw echo data of conjugate information.
38649 仿真结果表明,与经典MUSIC算法相比,改进的MUSIC算法参数估计精度更高,雷达散射截面重构拟合程度更好,且运算量增加不大,可有效提取出隐身目标的散射中心。 The simulation results show that compared with the classical MUSIC algorithm, the improved MUSIC algorithm parameters estimation precision is higher, the radar scattering cross section reconstruction better fitting degree, and the computational complexity increases, which can effectively extract the stealth target scattering center.
38650 针对目前健康因子构建方法存在的单调性和趋势性不够理想的问题,提出一种基于多尺度AlexNet网络的轴承健康因子构建方法。 Health factor construction method at the present time the monotonicity and trend, the problem of insufficient based on multi-scale AlexNet network is put forward a method of establishing bearing health factor.
38651 该方法首先利用连续小波分析将原始振动加速度信号转换为时频图,将时频图作为输入对多尺度AlexNet网络进行训练; This method firstly, the continuous wavelet analysis is used to change the original vibration acceleration signal conversion too frequency chart, the time-frequency diagram as the input of multi-scale AlexNet network was trained;
38652 然后利用训练好的网络在线构建测试轴承健康因子;最后根据健康因子评估准则评估初步构建的健康因子,利用评估结果调整网络参数,实现迭代优化,进一步提高健康因子的单调性和趋势性。 Then using the trained network online build test bearing health factor;Finally according to the factor of health assessment criteria to evaluate the preliminary build health factor, the use of evaluation results adjust the network parameters, implementation of optimization, to improve the monotonicity of health factors and trend.
38653 实验对比分析结果表明:该方法显著提高了健康因子的单调性与趋势性,不需要进行特征提取、特征选择、特征融合等步骤,具有较高的构建效率和泛化性。 Experimental comparison and analysis results show that the method significantly improves the monotonicity of health factors and trend, don't need to do a feature extraction, feature selection and feature fusion steps, high construction efficiency and generalization.
38654 针对单一智能体在导航过程中存在全球导航卫星系统(global navigation satellite system,GNSS)易受遮挡或干扰,惯性导航存在误差累积的问题,提出基于视觉的分层即时定位与地图构建(simultaneous localization and mapping,SLAM)空地多智能体协同算法。 For single agent exist in the process of navigation global navigation satellite system (global navigation satellite system, GNSS) vulnerable to keep out or interference, inertial navigation error accumulated problems, put forward based on visual hierarchical real-time localization and map building (simultaneous localization and mapping, SLAM) space more intelligent TiXie with algorithm.
38655 通过建立系统模型,采用基于扩展卡尔曼滤波融合欧氏点、逆深度点、锚定同质点3种不同特征点的分层SLAM算法,实现了对导航系统的辅助和增强。 By establishing the system model, USES the Euclidean based on extended kalman filtering fusion point, inverse depth point, three different anchor with particles feature points stratified SLAM algorithm, implements the auxiliary and increase of navigation system.