摘要
基于灰色理论和自适应数据融合技术的研究,提出一种基于自适应数据融合的新型灰预测GM(1,1)模型,并对整个建模过程进行了理论推导。该方法利用自适应数据融合以及累加再生成操作来提高非平稳时间序列的光滑度,从而减少样本序列的随机性,提高重构背景值的精确性以及灰预测GM(1,1)模型的拟合精度和预测精度。最后通过该方法对液浮陀螺仪零漂进行建模仿真,结果表明该方法辨识精度高,优于一般平均值法和灰预测方法,具有良好的应用价值。
A new modeling method of GM(1,1) was proposed based on the research of grey theory and adaptive data fusion. The whole modeling procedure of this method was established. By using adaptive data fusion and accumulated generating operation to smooth the original data sequence, the random characteristic of some non-stationary time series could be reduced, and the background value reconstructed is by far more precise. Moreover, the new GM(1,1) model’s fitting precision and prediction precision was improved. Experiments on identifying the liquid floated gyro signal zero drift show its precision is better than those of common grey forecasting methods. So the new method gives great application value.
出处
《系统仿真学报》
CAS
CSCD
北大核心
2006年第z2期62-65,共4页
Journal of System Simulation
基金
上海市重点学科(T0103)
上海市科委重点基础研究项目(04JC14038)
高等学校博士学科点专项科研基金(20040280017)