论文部分内容阅读
针对外弹道测量数据的有限性使随机误差的分布估计不够准确的问题,提出了基于Bootstrap(自助)方法的外弹道测量数据随机误差分布特性估计。首先对测量数据采用样条分频技术得到随机误差的初步估计,然后采用三次多项式拟合修正样本经验分布函数,替换传统Bootstrap方法中经验分布函数,得到随机误差的Bootstrap样本,计算Bootstrap样本的统计量,从而得到精度较高的外弹道测量数据随机误差模型。从仿真数据验证结果来看,解决了传统Bootstrap方法仿真随机样本相对比较集中的问题。
Aiming at the problem that the estimation of random error distribution is not accurate enough due to the limited external ballistic measurement data, the random error distribution estimation of the external ballistic measurement data based on Bootstrap method is proposed. Firstly, the preliminary estimation of random error was obtained by using the spline crossover technique. Then the cubic empirical polynomial fitting was used to modify the empirical distribution function of the sample to replace the empirical distribution function in the traditional Bootstrap method. The random error Bootstrap sample was obtained and the statistics of the Bootstrap sample Thus obtaining a more accurate random error model of the external ballistics data. From the verification results of simulation data, it solves the problem that the traditional Bootstrap method simulates the relatively concentrated random samples.