A new deep learning model based on improved alexnet for radiation source target recognition
Xiong Xu, Chenggang Wang · 2018
For the need of accurate radiation source target recognition, incorporating machine learning technology represented by deep learning, this paper explores a new mechanism and methods of intelligent learning which considering the characteristics of emitter signal data. This article puts forward an improved AlexNet as feature extractor, which achieved solidifying of fine feature extraction, and formed the intelligent recognition network model. With Automatic dependent surveillance-Broadcast (ADS-B) signal as the experimental object, 10 ADS-B pulse signals were collected as target data in the field of the airport as the training and testing samples for target recognition. The experiment use AlexNet and other neural network models to verify the effectiveness of the algorithm. The results show that the improved AlexNet network has faster training time and the comprehensive recognition rate is 98.28%.