中国惯性技术学报
中國慣性技術學報
중국관성기술학보
JOURNAL OF CHINESE INERTIAL TECHNOLOGY
2013年
3期
411-414
,共4页
党宏涛%杜祖良%任宏文%王常虹
黨宏濤%杜祖良%任宏文%王常虹
당굉도%두조량%임굉문%왕상홍
Bootstrap 方法%平台惯导系统%重复性%区间预测%性能评判
Bootstrap 方法%平檯慣導繫統%重複性%區間預測%性能評判
Bootstrap 방법%평태관도계통%중복성%구간예측%성능평판
bootstrap method%platform inertial navigation system%repeatability%interval prediction%performance evaluation
为了对存贮的平台惯导系统进行选择使用和性能评判,需要对每个惯导系统标定参数重复性进行当前或未来时间点的区间预测。考虑到每个惯导系统历次标定参数重复性具有不同参数的分布特性和批次,惯导系统在相同的存贮寿命阶段标定参数重复性具有相同参数的分布特性,提出了基于Bootstrap 方法的加权组合区间预测方法,并给出了1σ预测区间的上、下界均值。将该方法应用到不同批次的平台惯导系统加速计零偏重复性区间预测中,对应1σ的预测区间平均正确率为71.88%,表明该方法不仅具有良好的短期区间预测效果,而且依据预测区间的长度对平台惯导重复性性能评判也具有很好的应用价值。
為瞭對存貯的平檯慣導繫統進行選擇使用和性能評判,需要對每箇慣導繫統標定參數重複性進行噹前或未來時間點的區間預測。攷慮到每箇慣導繫統歷次標定參數重複性具有不同參數的分佈特性和批次,慣導繫統在相同的存貯壽命階段標定參數重複性具有相同參數的分佈特性,提齣瞭基于Bootstrap 方法的加權組閤區間預測方法,併給齣瞭1σ預測區間的上、下界均值。將該方法應用到不同批次的平檯慣導繫統加速計零偏重複性區間預測中,對應1σ的預測區間平均正確率為71.88%,錶明該方法不僅具有良好的短期區間預測效果,而且依據預測區間的長度對平檯慣導重複性性能評判也具有很好的應用價值。
위료대존저적평태관도계통진행선택사용화성능평판,수요대매개관도계통표정삼수중복성진행당전혹미래시간점적구간예측。고필도매개관도계통력차표정삼수중복성구유불동삼수적분포특성화비차,관도계통재상동적존저수명계단표정삼수중복성구유상동삼수적분포특성,제출료기우Bootstrap 방법적가권조합구간예측방법,병급출료1σ예측구간적상、하계균치。장해방법응용도불동비차적평태관도계통가속계령편중복성구간예측중,대응1σ적예측구간평균정학솔위71.88%,표명해방법불부구유량호적단기구간예측효과,이차의거예측구간적장도대평태관도중복성성능평판야구유흔호적응용개치。
In order to choose platform inertial navigation systems and evaluate their performances, the repeatability interval prediction should be made at current or future time points for each calibrated parameter of inertial navigation system. Considering that the calibrated parameter’s repeatability of each inertial navigation system has different parameter distribution characteristics, and the repeatability of different batch of inertial navigation systems in the same storage life stage shares the same distribution characteristics, this paper proposes a weighted combination interval prediction method based on bootstrap method, and the average values of 1σ at the upper and lower bound of the prediction intervals are given as well. By applying the method in the repeatability interval prediction of accelerometer’s calibrated bias in different batch of inertial navigation systems, the average accuracy of prediction interval 1σ is up to 71.88%, which shows that this method has good effects in short-range prediction, and has proper application value in repeatability performance evaluation of inertial navigation system according to the prediction range length.