ID 原文 译文
55407 具有攻击角度约束的导引律能够使导弹按指定方向打击目标,具有攻击时间约束的导引律能够使导弹按指定时间打击目标。 With attack Angle constraints of guidance law can make the missile target in the specified direction, has the guidance law can make the missile attack time constraints according to the specified target time.
55408 相比于传统的导引方法(如比例导引律),此类导引律均带来额外的机动,从而可能导致目标超出导引头视场(即使对于静止目标)。 Compared to traditional methods, such as proportional guidance law, the guidance of such guidance law are additional motor, which could lead to goal beyond the seeker field of view (even for stationary target).
55409 针对该问题,利用余弦函数的非线性特性,首先构造了一种带有导引头视场限制及攻击角度约束偏置项的比例导引律,并推导了该导引律的剩余时间估计。 Aiming at this problem, using the nonlinear characteristics of the cosine function, first of all, to construct a seeker with view limit and attack Angle constraint offset proportional guidance law, and the rest of the guidance law was deduced.
55410 然后通过附加剩余时间误差反馈项,得到了一种带导引头视场限制并同时具有攻击角度约束与攻击时间约束的导引律。 Then by attaching the rest of the time error feedback, got a band seeker field limits and at the same time with the constraint of the attack Angle and the attack time constraint guidance law.
55411 仿真结果验证了所提方法的有效性。 The simulation results verify the effectiveness of the proposed method.
55412 得益于隐层节点学习参数的随机选择,极限学习机(extreme learning machine,ELM)在学习速度极快的基础上,可以达到较为良好的分类性能。 Benefit from learning parameter of hidden layer nodes randomly selected, extreme learning machine (extreme learning machine, ELM) on the basis of learning speed, can reach a relatively good classification performance.
55413 但是,当隐层节点参数完全随机选择时,ELM的性能并不总能达到最优。 But when the parameters of hidden layer nodes randomly selected in full, the performance of ELM is not always able to achieve the optimal.
55414 本文提出多隐层输出矩阵极限学习机(multiple hidden layer output matrices extreme learning machine,M-ELM)方法解决这一问题,该方法通过对不同输出矩阵加权运算以优化隐层节点结构,其中权系数与输出权值在学习过程中同时分析确定。 In this paper, many hidden layer and output matrix extreme learning machine (multiple hidden layer the output matrices extreme learning machine, M - ELM) method to solve this problem, the method based on weighted arithmetic to optimize the hidden layer nodes of different output matrix structure, the weight coefficient and output weights at the same time in the learning process analysis to determine.
55415 另外,利用该方法可以实现特征级融合ELM。 In addition, the feature level fusion method can realize the ELM.
55416 实验证明,对于真实分类问题,M-ELM可以提供比ELM更为准确的分类结果。 Experiments show that for true classification problem, the M - ELM can provide more accurate classification result than ELM.