乏信息空间机械臂随机振动信号的灰自助评估

经典的统计学方法无法解决乏信息数据的评估问题。结合自助法和灰色系统理论,提出一种实现乏信息空间机械臂随机振动数据估计的灰自助方法。运用自助法对乏信息振动功率谱密度进行自助再抽样得到大量样本数据;利用灰色系统理论和最大熵理论建立灰自助模型,构建振动功率谱密度在不同频率点的灰自助分布。利用灰自助方法得到随机振动功率谱密度的真值估计和区间估计。提出了可靠度偏差和区间准确度2个指标对区间估计进行评价。灰自助方法与灰色方法和自助法的对比与测量实例表明,真值估计平均相对误差小于5%,在不同置信度水平下区间估计的准确度高于97%。 

Estimation of space manipulator random vibration signals with poor information based on grey bootstrap method

The classic statistical methods based on statistical theory can not solve the estimation of poor information data. A novel method based on grey bootstrap model for space manipulator random vibration signals with poor information was presented. The random vibration measurement data was processed by the bootstrap sampling．The bootstrap sequence was derived from the bootstrap distribution, and then the system theory was used to establish a bootstrap model. The true value and the interval of the space manipulator random vibration signals with poor information were estimated．grey bootstrap method is compared with grey method and bootstrap method. Experimental results show that grey bootstrap method has high accuracy. The validity of the proposed method is examined．The mean value of estimated relative errors are less than 5%, meanwhile, the estimated interval accuracy calculated with different confidence levels are more than 97%.