LSTM 모델을 이용한 대류층 천정 습윤지연 예측 기법
Prediction Method of Zenith Wet Tropospheric Delay with LSTM Model
1
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
Global navigation satellite system (GNSS) signal delay caused by the troposphere is a significant source of error for precision positioning. To mitigate this tropospheric delay effect, a precise point positioning (PPP) method estimates the tropospheric delay with carrier phase ambiguities. With accurately predicted tropospheric delay, this delay effect can be effectively minimized or the tropospheric delay estimation process can be improved. An artificial intelligence technique, Long Short-Term Memory (LSTM), was developed to predict the wet tropospheric delay using meteorological observations such as temperature, pressure, and relative humidity at GNSS monitoring stations. Five years of meteorological observation data from four GNSS measuring stations in the Korea were used to train the LSTM network model, which predicts one year of wet tropospheric delays at those stations. The predicted wet delays are compared with the wet delays estimated by the PPP method, and the forecast accuracy and geographical error differences are discussed.
Abstract
Global navigation satellite system (GNSS) signal delay caused by the troposphere is a significant source of error for precision positioning. To mitigate this tropospheric delay effect, a precise point positioning (PPP) method estimates the tropospheric delay with carrier phase ambiguities. With accurately predicted tropospheric delay, this delay effect can be effectively minimized or the tropospheric delay estimation process can be improved. An artificial intelligence technique, Long Short-Term Memory (LSTM), was developed to predict the wet tropospheric delay using meteorological observations such as temperature, pressure, and relative humidity at GNSS monitoring stations. Five years of meteorological observation data from four GNSS measuring stations in the Korea were used to train the LSTM network model, which predicts one year of wet tropospheric delays at those stations. The predicted wet delays are compared with the wet delays estimated by the PPP method, and the forecast accuracy and geographical error differences are discussed.