国防科技大学学报
國防科技大學學報
국방과기대학학보
JOURNAL OF NATIONAL UNIVERSITY OF DEFENSE TECHNOLOGY
2013年
6期
185-190
,共6页
李志猛%谈群%刘刚%郭少阳
李誌猛%談群%劉剛%郭少暘
리지맹%담군%류강%곽소양
卫星对地观测%任务完成概率%Logistic 回归模型
衛星對地觀測%任務完成概率%Logistic 迴歸模型
위성대지관측%임무완성개솔%Logistic 회귀모형
satellite observing task%enforceability probability%logistic regression method
针对及时响应用户提出的观测需求和高效实施卫星任务规划调度的需要,提出快速估计卫星对地观测任务完成可能性的现实问题,利用 Logistic 回归方法建立了一种卫星对地观测任务完成概率的估计模型。分析影响侦察任务完成概率的相关因素,包括侦察任务自身属性、资源约束、任务之间竞争关系等;从影响因素中提炼出刻画影响因素的模型变量,给出其量化方法,并初步构建了 Logistic 回归分析模型;基于卫星成像侦察任务规划系统开展实验,获取自变量与规划结果方面的样本数据;采用 Logistic 回归分析方法对样本数据进行分析,确定模型中的变量、参数及函数形式。结论表明,模型具有很好的统计特性。
針對及時響應用戶提齣的觀測需求和高效實施衛星任務規劃調度的需要,提齣快速估計衛星對地觀測任務完成可能性的現實問題,利用 Logistic 迴歸方法建立瞭一種衛星對地觀測任務完成概率的估計模型。分析影響偵察任務完成概率的相關因素,包括偵察任務自身屬性、資源約束、任務之間競爭關繫等;從影響因素中提煉齣刻畫影響因素的模型變量,給齣其量化方法,併初步構建瞭 Logistic 迴歸分析模型;基于衛星成像偵察任務規劃繫統開展實驗,穫取自變量與規劃結果方麵的樣本數據;採用 Logistic 迴歸分析方法對樣本數據進行分析,確定模型中的變量、參數及函數形式。結論錶明,模型具有很好的統計特性。
침대급시향응용호제출적관측수구화고효실시위성임무규화조도적수요,제출쾌속고계위성대지관측임무완성가능성적현실문제,이용 Logistic 회귀방법건립료일충위성대지관측임무완성개솔적고계모형。분석영향정찰임무완성개솔적상관인소,포괄정찰임무자신속성、자원약속、임무지간경쟁관계등;종영향인소중제련출각화영향인소적모형변량,급출기양화방법,병초보구건료 Logistic 회귀분석모형;기우위성성상정찰임무규화계통개전실험,획취자변량여규화결과방면적양본수거;채용 Logistic 회귀분석방법대양본수거진행분석,학정모형중적변량、삼수급함수형식。결론표명,모형구유흔호적통계특성。
The enforceability probability of satellite observation requires timely and accurate estimation to provide the fundamental basis for satellite task planning and scheduling.An estimation model of the enforceability probability of satellite earth observation is established based on the Logistic regression method.Firstly,the influence factors of observation enforceability probability were analyzed,including the property of observation task,resources limitation,competitive relations among tasks,etc.Secondly,independent variables were extracted and their quantitative methods were given,then the Logistic regression model was established.Thirdly,experiments were implemented on the planning and scheduling system of satellite imaging observation to obtain the sample data of independent variables and planning results.Moreover,the sample data was analyzed by using the Logistic regression method,and then the variables and parameters of the model and the function forms were all determined,Lastly,the effectiveness of the model was validated.