In this study, we propose a new method with an artificial intelligence infrastructure to estimate polarization of a radar signal. In modern electronic warfare and radar systems polarization of the return signal now poses essential information. Subspace-based algorithms such as MUSIC and ESPRIT have high computational costs for estimating polarization. Computational cost and performance analysis of the proposed method is conducted via simulations and results are discussed along with the existing solutions such as MUSIC algorithm. Simulation results show that the proposed algorithm reduce the computational cost compared to classical MUSIC. The proposed algorithm also reduce the polarization estimation error in low SNR scenarios.
A Fast Polarization Estimation Method with Convolutional Neural Networks
2022-06-27
1646807 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Ensemble convolutional neural networks for pose estimation
British Library Online Contents | 2018
|Springer Verlag | 2017
|CONVOLUTIONAL NEURAL NETWORKS FOR NIGHT-TIME ANIMAL ORIENTATION ESTIMATION
British Library Conference Proceedings | 2013
|