Application of Neural Network to GNSS-R Wind Speed Retrieval

This paper applies a machine learning (ML) algorithm based on the multi-hidden layer neural network (MHL-NN) for ocean surface wind speed estimation using global navigation satellite system (GNSS) reflection measurements. Unlike conventional wind speed retrieval methods that often depend on limited...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on geoscience and remote sensing 2019-12, Vol.57 (12), p.9756-9766
Main Authors: Liu, Yunxiang, Collett, Ian, Morton, Y. Jade
Format: Article
Language:English
Subjects:
Advanced Scatterometer (ASCAT)
Algorithms
Computer simulation
cyclone global navigation satellite system (CYGNSS)
deep learning
delay-Doppler map (DDM)
Doppler sonar
Global navigation satellite system
GNSS-reflectometry (GNSS-R)
Learning algorithms
Machine learning
Methods
multi-hidden layer neural network (MHL-NN)
Navigation
Navigational satellites
Neural networks
Neurons
Ocean surface
Retrieval
Root-mean-square errors
Satellites
Sea measurements
Sea surface
spaceborne remote sensing
Surface wind
Wind measurement
Wind speed
wind speed retrieval
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This paper applies a machine learning (ML) algorithm based on the multi-hidden layer neural network (MHL-NN) for ocean surface wind speed estimation using global navigation satellite system (GNSS) reflection measurements. Unlike conventional wind speed retrieval methods that often depend on limited scalar delay-Doppler map (DDM) observables, the proposed MHL-NN makes use of information captured by the entire DDM. Both simulated and real data sets are used to train and evaluate the performance of the MHL-NN and compare it to a conventional wind speed retrieval method and other prevailing ML algorithms. The results show that the MHL-NN algorithm outperforms the other methods in terms of the root mean square error (RMSE) and mean absolute percentage error (MAPE) of the wind speed estimation.
ISSN:0196-2892
1558-0644
DOI:10.1109/TGRS.2019.2929002