融合仿真数据的小子样圆概率误差评估

针对应用Bayes方法评估圆概率误差（CEP）时仿真数据容易“淹没”现场实弹数据，导致评估结果无效甚至错误的问题，提出一种基于“有效样本”的飞行与仿真数据融合算法与融合仿真数据的小子样CEP评估方法。使用重要性抽样方法中抽样效率的评估模型度量仿真先验分布与总体分布之间的偏差，建立考虑仿真先验信息有偏差时仿真先验分布的确定方法。该方法在融合评估时仿真样本权重同时依据仿真先验分布与总体分布之间的偏差和样本量调整，而不是原方法中仅仅依赖验前试验的试验次数分配。实验和应用案例表明，新方法的估计性能显著好于传统方法，可以有效解决装备鉴定和定型试验中小子样CEP评估问题。

Small Sample Circular Error Probable Estimation Based on Fusing Simulation and Flight Test Data

When Bayes method is used to evaluate the circular error probable (CEP), the simulation data is easy to “submerge” the live ammunition flight test data, which often leads to the ineffectiveness or even mistake of evaluated results. To solve this problem, a flight and simulation data fusion algorithm based on “effective sample” and a small sample CEP evaluation method for fusing simulation data and flight data are proposed. The evaluation model of sampling efficiency in importance sampling method is used to measure the deviation between the simulation prior distribution and the overall distribution, and the simulation prior distribution considering the deviation of simulation prior information is established. In the process of fusion evaluation, the weights of simulation samples are calculated according to the deviation between the prior distribution and the overall distribution and the sample size, instead of the original method only relying on the test times of prior test. Experiments and application cases show that the proposed method is significantly better than the traditional method in CEP estimation, which can effectively solve the problem of small sample CEP evaluation in equipment evaluation and validation test.