基于自适应小波SVR算法的陀螺漂移预测  

Gyroscope Drifting Prediction Based on Adaptive Wavelet SVR

作　　者：蔡艳宁[1] 曹小平[2] 胡昌华[1] 

机构地区：[1]第二炮兵工程学院302教研室,陕西西安710025 [2]国防科技大学机电工程学院与自动化学院,湖南长沙410073

出　　处：《控制工程》2007年第S3期26-29,共4页Control Engineering of China

摘　　要：针对某导弹的陀螺漂移趋势预测问题,提出了一种自适应小波支持向量回归机算法。该算法通过分析加入到训练集中的样本是否为支持向量以确定新样本的预测方法。此外,该模型将小波核函数引入其中以提高算法的泛化能力。文中给出了自适应SVR算法的具体步骤并将其用于预测某导弹的陀螺漂移趋势。仿真结果表明,该算法能够获得较好的预测精度。An adaptive wavelet support vector regression algorithm is proposed for the prediction of the gyroscope drifting tendency.The algo- rithm selects the prediction method of the new samples by anglicizing if the current sample added to the training set is support vector.In addit- ion,the wavelet kernel function is introduced to improve the generalization ability of the model.The algorithm is presented in detail and used to predict the missile gyroscope drifting tendency.The simulation results show the model can attain good prediction precision.

关 键 词：支持向量回归机 小波核 陀螺漂移 预测 

分 类 号：TN966[电子电信—信号与信息处理]