A support vector machine‐based VVP wind retrieval method

Abstract

Velocity volume processing (VVP) is a common wind retrieval method for single-Doppler radar. However, hindered by the problem of the ill-conditioned coefficient matrix, VVP method has not been applied effectively. In this paper, a support vector machine-based VVP (SVM VVP) method is proposed to improve the VVP method. SVM VVP retrieves wind field through constructing a support vector regression machine for the radial velocity. The new method is applied to a real case, and the retrieval shows that SVM VVP can give accurate and high-resolution results. It is helpful for studying wind fields of small-scale precipitation systems.

A Spatio-Temporal Neural Network for Fine-Scale Wind Field Nowcasting Based on Lidar Observation

“…Manuscript received ; revised field. To estimate the wind vectors in specific areas, such as the aircraft approach and take-off corridors [16], retrieval methods have been widely studied [17], [18], [19], [20], [21], [22].…”

An Improved Estimation and Gap-Filling Technique for Sea Surface Wind Speeds Using NARX Neural Networks

“…Li et al (2012) and Tokinaga and Xie (2011) have used a power series regression to estimate wind speeds from significant wave heights, while Zeng et al (2016) have used a neural network to estimate wind speed using different oceanographic measurements. Among the most recent wind speed estimation methods, Clarizia and Ruf (2017) proposed a Bayesian wind speed estimator using significant wave height information inferred from Global Navigation Satellite System reflectometry (GNSS-R) data, while Li et al (2015Li et al ( , 2017 proposed the use of support vector machines (SVM) to estimate wind speed from a single-Doppler radar.…”