Abstract:
The detection and classification o f r adar targets have become an important topic nowadays, and radar sensors play a key role in these operations because of their robust...Show MoreMetadata
Abstract:
The detection and classification o f r adar targets have become an important topic nowadays, and radar sensors play a key role in these operations because of their robustness to different weather and light conditions. In this paper, a classification a lgorithm u sing b oth o verlapped R D m ap (Range-Doppler map) method and GRU (Gated recurrent unit) based network is proposed. The overlapped method is based on the using information of both Doppler signature and spatial size of target. Moreover, due to computational requirements and the usage of relatively small data sets in radar applications, a simpler LSTM (Long short-term memory) variant, which is GRUs, is proposed. The simulations are designed and performed by using MATLAB 2022A and its Deep Learning Toolbox. The experimental results obtained are proposed, with an increase of 9.05 % in helicopter classification i n R adar A a nd 3 4.27 % in Radar B is achieved.
Published in: 2022 30th Telecommunications Forum (TELFOR)
Date of Conference: 15-16 November 2022
Date Added to IEEE Xplore: 22 December 2022
ISBN Information: