ID 原文 译文
11314 针对应用于受不确定性干扰和噪声影响的卫星自主导航系统中的无迹卡尔曼滤波(unscented Kalman filter,UKF)存在估计精度低、跟踪性能差和鲁棒性弱等缺陷,提出一种改进的强跟踪平方根UKF(strong tracking square-root UKF,STSRUKF)导航方法。 Against interference and noise is applied to the uncertainty of satellite autonomous navigation system no trace of the Kalman filter (unscented Kalman filter, UKF) low estimation precision, the tracking performance is poor and weak defects such as robustness, proposed a modified strong tracking square root UKF (strong tracking the square - root UKF, STSRUKF) navigation methods.
11315 该方法中利用星敏感器和光学导航相机设计出导航方案,并通过转换方程将间接量测量转换为观测量。 ‭The method of using star sensor and optical navigation camera designed navigation scheme, and through the transformation equation transform indirect measurement for observation.
11316 针对平方根UKF(square-root UKF,SRUKF)在高阶系统中因为sigma点的零权值系数是负的或者数值计算误差太大时而可能造成滤波器发散问题,采用一种改良的平方根分解方法,改善了滤波器的稳定性。 ‭In view of the square root UKF (square - root UKF, SRUKF) in a higher order system for sigma point is negative or zero weight coefficients of numerical calculation error is too big and sometimes may cause the filter divergence problem, using a modified square root decomposition method, the stability of the filter is improved.
11317 同时,基于强跟踪滤波器理论(strong tracking filters,STF),引入多重自适应衰减因子调节协方差矩阵,使得滤波器具有强跟踪能力和克服系统模型不确定的鲁棒性,改善了滤波器的估计精度。 ‭At the same time, based on the theory of strong tracking filter (strong tracking filters, STF), the introduction of multiple adaptive damping factor to adjust the covariance matrix, the filter has strong tracking ability and overcome the robustness of uncertain system model, the estimation precision of the filter is improved.
11318 将该方法应用于卫星自主导航系统中,实验仿真结果表明,相对于平方根UKF和STF,该方法不仅保证了系统的可靠性,还提高系统的导航精度和改善系统的鲁棒性及跟踪能力。 ‭The method was applied to satellite autonomous navigation system, the experimental simulation results show that relative to the root UKF and STF, this method not only ensure the reliability of the system, and improve the navigation precision of the system and improve system robustness and tracking ability.
11319 针对传统的ε-不敏感支持向量回归机(ε-insensitive support vector regression,ε-SVR)未充分考虑局部支持向量对回归预测结果的影响,不利于提高回归预测精度的问题,提出了一种ε-SVR预测误差校正方法。 Is insensitive to the epsilon - against the traditional support vector regression machine (epsilon insensitive - support vector regression, epsilon - SVR) not fully consider the influence of the local support vector regression prediction results, is not conducive to improve the predictive accuracy of regression problem, this paper proposes a epsilon - SVR prediction error correction methods.
11320 该方法以期望预测值与ε-SVR回归预测值及局部支持向量间的欧氏距离和最小为目标函数。 The method to forecast and epsilon - SVR regression prediction and the Euclidean distance between local support vector and minimum as the objective function.
11321 以ε不敏感损失带(ε-tube)宽度为约束条件,通过利用高维特征空间中ε-tube边界上和边界外的局部支持向量对ε-SVR的回归预测值进行误差校正。 The loss is not sensitive to epsilon band width (epsilon - tube) as constraint conditions, by using the epsilon - tube in high dimensional feature space outside the boundary and boundary of local support vector regression prediction of epsilon - SVR error correction.
11322 利用人工产生的不同分布数据集和UCI数据集进行的仿真结果表明,与传统的ε-SVR相比,该文方法具有更高的预测精度和更强的泛化能力。 Different distribution produced using artificial data set and UCI data sets of the simulation results show that compared with the traditional epsilon - the SVR, this method has higher prediction precision and better generalization ability.
11323 异构无线传感器网络(heterogeneous wireless sensor works,HWSN)能有效降低数据转发延迟、网络能量消耗,是一种更现实的网络模型,基于HWSN的k容错性拓扑控制是一类NP-难问题。 Heterogeneous wireless sensor networks (heterogeneous wireless sensor works, HWSN) can effectively reduce the traffic delay, the network energy consumption, is a more realistic network model, based on HWSN k fault tolerant topology control is a kind of NP - hard problem.