CNN-32DC: An improved radar-based drone recognition system based on Convolutional Neural Network
Ann Janeth Garcia, Ali Aouto, Jae‐Min Lee, Dong‐Seong Kim
Abstract
This paper proposes a system that will guard infrastructures against incoming threats from drones by detecting it with the use of a radar device-based detection scheme. The database acquired, named Real Doppler RAD-DAR (Radar with Digital Array Receiver) is constructed by a Microwave and Radar Group. The radar used uses a Frequency Modulated Continuous Wave (FMCW) on an 8.75 GHz based frequency band with a BWmax of 500 MHz. The proposed Convolutional Neural Network (CNN), CNN-32DC is varied with different number of filters, combination layers, and number of feature extraction blocks, the preference that will give the most accurate result was selected and compared with different machine learning and classification learning algorithms gained an accuracy that exceeds other networks with less processing time.