Fingerprint Extraction and Classification of Wireless Channels Based on Deep Convolutional Neural Networks

Abstract

We propose the use of a deep convolutional neural network (DCNN) for fingerprint feature extraction and classification of wireless channels based on software defined radio. In the past, conventional classification schemes for wireless channels rely heavily on artificial extracting features, which limit their scalability. In this letter, we solve this problem based on DCNN and spectrogram. DCNN can automatically learn features and conduct classification using the gathered data. Our approach is tested in real-life environment. From the experiment, our DCNN model can extract the fingerprint features of wireless channels effectively. At the same time, it shows 96.46% accuracy for wireless channel classification.