ID	原文	译文
6454	然后分别采用了Wiener过程的基本模型和3个具有随机参数的Wiener过程模型进行退化建模,运用贝叶斯方法进行参数估计,并通过模型验证、分位数图(quantile-quantile plot,QQ图)和改进的留一法交叉验证等模型选择方法综合判断确定最优模型。	Then respectively adopted the Wiener process, the basic model and three Wiener process model with stochastic parameters of degradation model, using the bayesian method for parameter estimation, and through the model validation, the quantile chart (quantile - quantile plot, QQ photo) and improve the comprehensive judgment method of cross validation model selection method to determine the optimal model.
6455	最后采用裂纹增长数据实例验证了所提方法的有效性。	Finally using crack growth data to verify the effectiveness of the proposed method.
6456	针对多普勒雷达目标跟踪问题,提出了基于预测值量测转换的序贯滤波目标跟踪算法。	According to doppler radar target tracking problem, based on the predicted value measurement conversion of sequential filter target tracking algorithm.
6457	该算法采用基于预测信息的量测转换方式解决量测与目标运动状态的非线性。其中位置量测转换采用乘性去偏、伪量测转换采用加性去偏；	The algorithm is based on the prediction information measurement transformation method to solve measure and target motion state of the nonlinear, including position measurement conversion by take sex, pseudo measurement conversion by additive to partial；
6458	量测转换误差的统计特性基于预测值进行推导,并采用序贯滤波方式处理伪量测以实现目标跟踪。	measuring the statistical properties of a conversion error is derived based on the forecast, and USES the sequential filtering approach to pseudo measurement to realize target tracking.
6459	同时将该算法扩展到机动目标跟踪情况,综合利用了位置量测与伪量测滤波部分输出的概率组合作为该模型的更新概率,利用模型更新概率对各滤波器的状态估计结果进行加权求和获得最终估计。	At the same time to extend this algorithm to the situation of maneuvering target tracking, comprehensive utilization of the position measurement combined with the probability of pseudo measurement filter part of the output as the probability of the updating of the model, using the model to update the probability of each filter state estimation results are weighted sum estimates that in the end.
6460	仿真结果表明该算法与传统的序贯滤波跟踪算法相比,具有更高的跟踪精度,其扩展方法可实现有效的机动目标跟踪。	The simulation results show that the algorithm is compared with the traditional tracking algorithm of sequential filtering, has higher tracking precision, the extension methods can achieve effective maneuvering target tracking.
6461	针对目前随机过程退化模型错误指定研究较少,且主要集中在线性模型中的现状,研究了逆高斯过程(inverse Gaussian,IG)的两类错误指定:	Aiming at random model error to specify the process of degradation is less, and are mainly concentrated in the linear model of the status quo, to study the inverse Gaussian process (inverse Gaussian, IG) of the two types of errors specified:
6462	不含随机效应情况下非平稳IG过程被错误指定为非线性Wiener过程,以及含随机效应IG过程被错误指定为简单IG过程。	excluding non-stationary random effects cases IG process have been designated as a nonlinear Wiener process, as well as containing random effects IG process have been designated as a simple process of IG.
6463	基于伪最大似然估计近似正态性理论获得了这两种情形下伪平均失效前时间(mean-time-to-failure,MTTF)估计的分布特征,并以某合金疲劳裂纹数据为例,分析比较了模型错误指定对MTTF的影响。	Based on the theory of pseudo approximate maximum likelihood estimate normality won both cases pseudo average time before failure (mean - time - to - failure, MTTF) estimate of the distribution characteristics, and with a metal fatigue crack data as an example, analyzed the influence of model error for MTTF specified.