ID 原文 译文
50097 在发射端，该算法通过预编码矩阵将信号扩展到所有子载波上，从而降低部分子载波深衰落对扩展前原始信号的影响。 In the transmitter, the algorithm by precoding matrix will be extended to all the subcarrier signal, thereby reducing parton carrier deep decline for extension before the influence of the original signal.
50098 在接收端，利用最小均方差误差排序QR分解(minimum mean square error sorted QR decomposition，MMSE-SQRD)软输入软输出干扰消除均衡算法，一方面避免传统基于最小均方误差(minimum mean square error，MMSE)并行软干扰消除均衡算法中复杂的矩阵求逆运算，进而降低了算法复杂度，另一方面利用信道排列优先检测信噪比最大的传输符号提高检测准确性。 At the receiving end, using the minimum mean square deviation error sorted QR decomposition (minimum mean square error sorted QR decomposition, MMSE - SQRD) soft input soft output interference elimination equalization algorithm, on the one hand, to avoid the traditional based on minimum mean square error (minimum mean square error, MMSE) parallel soft interference elimination equalization algorithm in the complex matrix inversion, which reduces the algorithm complexity, on the other hand using channel arrangement priority detection signal-to-noise ratio was improved maximum transmission symbol detection accuracy.
50099 同时通过预编码对重构信号中误差进行扩展，进而缓解在迭代干扰消除过程中的误差传播。 At the same time by precoding extensions error of reconstructing signal, thereby alleviating the error propagation in the process of iterative interference elimination.
50100 仿真结果证明，在2发2收场景下，误码率在10-3时，算法经过5次迭代后系统性能相比于现有的迭代均衡算法改善约4dB。 The simulation results show that in 2 2 scenarios, bit error rate in 10-3, after 5 iteration algorithm iterative equalization algorithm to improve the system performance compared with the existing about 4 db.
50101 针对具有多维状态变量、多种工作模式和故障模式的复杂工程系统，提出一种基于综合健康指数(synthesized health index，SHI)与相关向量机(relevance vector machine，RVM)的系统级失效预测方法。 For a multidimensional state variables, a variety of complex engineering system, working mode and fault mode is put forward based on the comprehensive health index (synthesized the health index, SHI) and relevance vector machine (relevance vector machine, RVM) system level failure prediction method.
50102 在离线训练阶段，先根据有限失效历史数据建立各工作模式下的健康评估模型，并据此获得各历史退化轨迹的SHI序列; In offline training phase, based on historical data limited failure to establish the working mode of health assessment model, and on the basis of the historical trajectory of degradation SHI sequence;
50103 然后再使用RVM对这些序列进行回归处理，进而辨识出与回归曲线最为匹配的函数模型。 Then using RVM to return to deal with these sequences, and then identify and regression curve match most function model.
50104 最后，外推预测出系统剩余使用寿命的概率密度分布。 Finally, the extrapolation to predict system probability density distribution of the remaining service life.
50105 该方法成功应用到涡轮发动机的失效预测案例。 The method is successfully applied to the turbine engine failure prediction.
50106 神经网络动态优化算法(neural dynamic optimization，NDO)的主要特点是能使神经网络逼近最优解。 In order to solve the break, the craft and craft when disturbed trajectory optimization problem, a new generation of reentry vehicle type need to have the ability of real-time online trajectory optimization.