基于EMD-LSTM耦合预测模型的BDS多路径误差削弱方法研究

BDS multipath errors reducing method based on EMD-LSTM coupled prediction model

摘要: 北斗卫星导航系统(BDS)在短基线测量中存在的多路径误差是影响定位精度的主要误差项. 针对多路径误差的非线性以及坐标序列的非平稳特性,拟采用经验模态分解(EMD)与长短期记忆网络(LSTM)结合的方法,构建EMD-LSTM耦合预测模型,对多路径误差进行预测,削弱多路径误差的影响. 实验结果表明,EMD-LSTM耦合预测模型能够有效地削弱了多路径误差影响,E、N、U方向精度分别提高了22％、36％、40％.

Abstract: The multipath errors in short baseline measurement of BeiDou Navigation Satellite System (BDS) were the main errors affecting positioning accuracy. Aiming at the nonlinearity of multipath errors and the non-stationarity of coordinate series, a coupled prediction model of EMD-LSTM was proposed by combining empirical mode decomposition (EMD) with long short-term memory (LSTM) to predict multipath errors and weaken the influence of multipath errors. The experimental results show that the EMD-LSTM coupled prediction model can effectively reduce the multipath errors, and the E, N,and U directions were respectively improved by 22％, 36％, and 40％.