摘要
为了提高遥测数据预测的精度和实时性,针对遥测数据的非平稳性和周期性特点,引入小波分析的预测技术,提出了一种对遥测数据序列进行不同频段上的分解方法:遥测数据时间序列依据选定的N阶多贝西小波和分解尺度值2分解为低频分量和高频分量,针对不同分量建立了基于马特拉算法、周期自回归模型和指数平滑法的时间序列短期预测模型,各分量预测结果经小波变换的逆算法重构后输出.仿真实验结果表明该方法满足遥测数据工程预测要求,能够有效地解决遥测数据的短期预测问题.通过对遥测数据短期预测结果的研究分析可提前判断卫星潜在的趋势,为指挥人员的正确决策提供科学依据.
To improve the accuracy and real-time of telemetry data prediction,according to the non-stationarity and periodicity of telemetry data,the wavelet analysis forecasting technology was introduced and a decomposition method of different frequency ranges on the sequence of telemetry data were proposed.The time series of telemetry data were decomposed into low frequency components and high frequency components based on dbN wavelet function and decomposition scale 2.A short-term prediction model based on Mallat algorithm,periodic auto-regression model and exponent-smoothing algorithm was established for different components.The last result was output after reconstruction by the inverse algorithm of wavelet transform.Simulation results show that such method meets the project requirements of telemetry data,which can solve the short-term forecasting problem of telemetry data effectively.Moreover,it can forecast the potential tendency in advance based on the analysis of short-term forecasting result,which can provide a scientific basis for commander in right decision-making.
出处
《武汉工程大学学报》
CAS
2014年第2期73-78,共6页
Journal of Wuhan Institute of Technology
基金
国家自然科学基金项目(61103143)
河南省教育厅科学技术研究重点项目(14B520014)
周口师范学院青年基金项目(zknuc0214)
关键词
马特拉算法
周期自回归模型
指数平滑法
短期预测
Mallat algorithm
periodic auto-regression model
exponential smoothing method
short-term forecasting
