认知不确定下动态试验误差条件抽样分析方法

Dynamic Experiment Error Analysis Method Based on Conditional Sampling Under Epistemic Uncertainty

为满足武器系统动态试验误差分析的需求,针对试验误差分析中设备和模型的信息缺失影响分析结果的问题,提出了认知不确定条件下动态试验误差条件抽样分析方法.该方法将试验系统的认知不确定性分为模型认知不确定性和测量设备认知不确定性,并推导了二者概率分布间的联系.设计了条件抽样方法以得到认知不确定性和试验误差分布关系.根据试验结果,得到最佳认知不确定参数和试验误差分布.使用该方法对某引信定距空炸试验系统进行了分析,得到了其误差分布,并通过与试验数据及忽略认知不确定条件时误差分布仿真结果的对比表明,运用该方法获得的引信定距空炸试验误差分布与试验结果的相符程度更高.

A dynamic experiment error（DEE）analysis method was proposed for analyzing the probability distribution of dynamic experiment error based on conditional sampling under epistemic uncertainty（EU）to improve the error analysis accuracy tampered with lack of experiment equipment and model information.Firstly,the EU of experiment system was divided into measure device EU and model EU,and the relationship between two kinds of their probability distribution was formulated.Then a conditional sampling method was designed to establish the relationship between DEE distribution and EU.The best DEE and EU parameters were gotten by comparing the sampled distribution with the experiment result.Finally,the method was validate in a fuze air burst distance experiment.Compared with experiment result and the error distribution simulation result without taking the epistemic uncertainty into account,the error distribution resulted in this method is more match with the experiment result.