一种WOA-RBF的BDS精密卫星钟差短期预报方法
A parameter optimized RBF method for short-term prediction of BDS precision satellite clock errors
李特 1杨振 2田静 1郭建春 3郑伟 1范舒畅4
作者信息
- 1. 黑龙江工程学院 测绘工程学院,哈尔滨 150050
- 2. 南方测绘科技股份有限公司,广州 510000
- 3. 黑龙江第一测绘工程院,哈尔滨 150500
- 4. 广东省交通规划设计研究院集团股份有限公司,广州 510507
- 折叠
摘要
针对当前北斗卫星导航系统(BDS)精密卫星钟差预测模型精度不高、预测误差较大等问题,提出一种将鲸鱼优化算法(WOA)和径向基函数神经网络(RBF)相结合的钟差预测模型—鲸鱼算法优化的RBF组合模型(WOA-RBF):利用四分位法和分段线性插值法完成数据预处理,通过鲸鱼优化算法对RBF中的扩展速度和输出层线性权重进行寻优,得到最优参数,最终得到优化后的输出结果.实验结果表明:与二次多项式(QP)模型、灰色模型(GM)和径向基函数神经网络(RBF)模型相比,WOA-RBF模型优势明显.在预报时长为6 h时,均方根误差在0.25 ns以内;在预报时长为12 h时,均方根误差在0.27 ns以内,证明了WOA-RBF模型在精密卫星钟差短期预报中的准确性和可行性.
Abstract
Aiming at the problems of low accuracy and large prediction errors in the current BeiDou navigation satellite system (BDS) precision satellite clock error prediction model,a clock error prediction model combining whale optimization algorithm (WOA) and radial basis function neural network (RBF) was proposed. Whale algorithm optimized RBF combination model (WOA-RBF) was that the quartile method and segmented linear interpolation method were used to complete data preprocessing. The whale optimization algorithm was used to optimize the expansion speed and linear weight of the output layer in RBF,obtain the optimal parameters,and finally obtain the optimized output result. The experimental results showed that compared with the quadratic polynomial (QP) model,grey model (GM),and radial basis neural network (RBF) model,the WOA-RBF model had significant advantages. When the forecast duration was 6 h,the root mean square error was within 0.25 ns;When the forecast duration was 12 h,the root mean square error was within 0.27 ns,which proved the accuracy and feasibility of the WOA-RBF model in short-term prediction of precision satellite clock errors.
关键词
北斗卫星导航系统(BDS)/钟差预报/鲸鱼优化算法/径向基函数神经网络Key words
BeiDou navigation satellite system (BDS)/clock error prediction/whale optimization algorithm (WOA)/radial basis function neural network引用本文复制引用
基金项目
黑龙江工程学院青年基金(2021QJ02)
出版年
2024
NETL
NSTL
万方数据