ID 原文 译文
57528 三维旋转不变子空间( 3D-ESPＲIT) 算法估计三维几何绕射理论( GTD) 的散射中心模型时，参数精度不高，噪声鲁棒性较差，为此，提出一种改进的 3D-ESPＲIT 算法. The noise robustness and parameter estimation performance of the classical three-dimensionalestimating signal parameter via rotational invariance techniques ( 3D-ESPＲIT) algorithm are poor whenestimating parameters of the three dimensional geometric theory of diffraction( 3D-GTD) model at low sig?nal-to-noise ratio ( SNＲ) .
57529 通过构建原始回波数据的共轭矩阵、对协方差矩阵叠加、取平均、平方处理，可提高对目标原始回波数据的利用率，并增大信号特征值与噪声特征值之间的差距，增大了信噪比，从而提高参数的估计精度. To solve this problem，a modified algorithm is proposed. The modified algo?rithm improves the parameter estimation accuracy by constructing conjugated matrix of the original back?scattered data，averaging the covariance matrices and performing quadratic power on the total covariancematrix to get another novel covariance matrix.
57530 基于不同算法估计得到的三维 GTD 模型参数，比较了不同算法估计的参数均方差. The modified algorithm can broaden the differences be?tween the eigenvalues of noises and signals，which is equivalent to increase SNＲ.
57531 仿真结果 表 明，改进算法的参数估计性能及噪声鲁棒性均优于经典 3D-ESPＲIT 算法和改进的免配对3D-ESPＲIT算法. Based on the parame?ters estimated by different algorithms，root-mean-square-errors of different parameters can be compared. Simulations verify that the parameter estimation performance and noise robustness of the modified algo?rithm are better than the classical 3D-ESPＲIT algorithm and the improved algorithm without pair.
57532 针对复杂接触环境下机器人柔顺装配困难的问题，提出了一种基于受力信息精准获取与改进力/位混合控制算法的柔顺装配方法. Aiming at the problem that the difficulty of compliant assembly of robot in complex contact en- vironments，a compliant assembly method based on accurate acquisition of force information and improved hybrid force /position control algorithm is presented.
57533 在对机器人末端力传感器进行零点补偿及负载重力补偿后，通过矩阵变换实时计算出装配接触点处机器人末端执行器的真实受力，并在力控制过程中让机器人的运动遵循双曲正切速度—力关系. In the method，after bias and load gravity compensa- tion are performed on the robot end force sensor，the real contact force of the robot end-effector at the as- sembly contact point is calculated in real time by matrix transformation，and the motion of the robot follows velocity-force relationship of the hyperbolic tangent during force control.
57534 实验结果表明，机器人在装配过程中并未发生明显的抖动，所提算法具有良好的柔顺装配效果. Experiments show that the robot did not shake significantly during the assembly process，and the proposed algorithm has a good compliant assembly effect.
57535 为提高云平台负载预测的精度，提出了一种基于时序相关性的多负载序列联合预测方法. A novel approach，joint prediction of multi-workload sequences，was proposed based on tempo- ral correlation.
57536 首先，为获得相似的负载序列，采用长短时记忆网络提取负载序列的时序特征，再利用层次聚类法，获得在时序特征空间相似的负载序列类; Firstly，long short-term memory was used to extract the temporal feature among workload sequences for obtaining the similar workload sequences，while hierarchical clustering algorithm was used to obtain workload sequences clusters.
57537 其次，对获得的每个负载序列类分别构建多任务学习模型，挖掘和利用负载序列间隐藏的共享领域知识，提高模型泛化能力和预测精度，并实现多个负载序列的联合预测. Then，construct multi-task learning model respectively for each ob- tained sequence clusters，capture and utilize the shared domain knowledge among multiple workload se- quences，so as to achieve joint prediction of multiple workload sequences as well as improve generalization ability and prediction accuracy of the model.