运载火箭精度评估的Bayes异质先验融合方法(英文)

Heterogeneous information fusion of Bayesian model with applications in accuracy evaluation of carrier rocket

为充分利用运载火箭观测中的不同观测空间和过程的信息来进行精度评估 ,针对该背景建立了异质先验融合的数学描述。研究了飞行试验中不同观测空间和过程的异质先验信息和数据 ,基于不同观测过程的解析关系 ,将间接过程的先验和观测数据算出的后验分布转换成落点观测空间上的先验 ,与原落点的先验进行了最大熵加权融合 ,得到混合后验分布 ,从而结合落点观测数据给出评定结果。在无法解算出精确的制导工具误差系数的情况下 ,这种方法充分利用了弹道跟踪数据、工具误差系数的地面测试先验值、落点先验及落点数据 ,稳健性更好 。

As far as the system observation for accuracy evaluation of carrier rocket is concerned,there may exist different types of observed data and priors.Heterogeneous information means that the different information describing the different characteristics of the same object.Since all of the information is relevant to the same object,the fusion is possible.It is a key problem that how to fuse the heterogeneous information to obtain the better evaluation result.Therefore,the different heterogeneous information and data is thoroughly studied,moreover,the mathematical description for information fusion of different parameter priors and data is constructed in this paper.Based on their relationship between different parameters,indirect prior and observation data is transformed into prior in impact point observation space,which is fused with original prior by weight determined by Maximum entropy rule to obtain the mixed posterior distribution.Therefore,the test results can be given by combining posterior distribution and impact error observed data.Then its application on evaluating guidance systematic error is elaborated as it applies trajectory tracking data,test value of coefficients of guidance instrumentation systematic error,impact point observation data and prior.Especially,the advantage of this method lies in its application in case that guidance instrumentation systematic error may not be computed precisely.Finally a detailed example on evaluation of carrier rocket is given to verify the theory.