振动与冲击
振動與遲擊
진동여충격
JOURNAL OF VIBRATION AND SHOCK
2015年
17期
137-141
,共5页
遥感卫星振动%支持向量机%时间序列%ARMA 模型%在线预测
遙感衛星振動%支持嚮量機%時間序列%ARMA 模型%在線預測
요감위성진동%지지향량궤%시간서렬%ARMA 모형%재선예측
remote sensing satellite vibration%SVM%time series%ARMA model%online prediction
将支持向量机算法与时间序列原理相结合，可构造出基于时间序列的支持向量机模型。通过对大量遥感卫星振动数据进行分析，得出该卫星振动规律为有随机波动成分的简谐振动。应用时间序列的原理，动态更新模型训练集和预测集，构建基于时序回归的支持向量机在线预测模型。模型测试结果表明，这种方法可以比较准确有效地实现振动趋势的提前预测，为振动抑制措施的快速实现提供帮助。
將支持嚮量機算法與時間序列原理相結閤，可構造齣基于時間序列的支持嚮量機模型。通過對大量遙感衛星振動數據進行分析，得齣該衛星振動規律為有隨機波動成分的簡諧振動。應用時間序列的原理，動態更新模型訓練集和預測集，構建基于時序迴歸的支持嚮量機在線預測模型。模型測試結果錶明，這種方法可以比較準確有效地實現振動趨勢的提前預測，為振動抑製措施的快速實現提供幫助。
장지지향량궤산법여시간서렬원리상결합，가구조출기우시간서렬적지지향량궤모형。통과대대량요감위성진동수거진행분석，득출해위성진동규률위유수궤파동성분적간해진동。응용시간서렬적원리，동태경신모형훈련집화예측집，구건기우시서회귀적지지향량궤재선예측모형。모형측시결과표명，저충방법가이비교준학유효지실현진동추세적제전예측，위진동억제조시적쾌속실현제공방조。
A support vector machine (SVM)model based on the time series principle can be constructed with the support vector machine algorithm and the time series principle.Through analyzing a large number of a remote sensing satellite's vibration data,it was shown that its vibration law is a simple harmonic vibration with some random wave components.Training set and prediction set of its model were dynamically updated by using the principle of time series, and then a SVMonline prediction model was built based on the time sequence regression.The model test results showed that this method can effectively predict a remote sensing satellite's vibration trends in advance,and give helps to realize rapid vibration suppression measures.
