Convolutional Neural Networks for Special State Identification Based on Transient Signal

Abstract

Abstract. Most exiting algorithm focused on the specific emitter identification, neglected the state identification of emitter. However, using the fingerprinting to identify and estimate the user behavior is useful in many legitimate scenes.While the emitter is moving or shaking, a certain amount of vibrations are generated, which can be investigated as the basis of state identification. However, the subtle vibrations are difficult to capture. In this paper, we propose a novel state analysis technique to identify the state of emitter based on transient signal. The transient signal from the target emitters contains import special information. In order to find the subtle difference, 1-D convolution neural networks (CNNs) is introduced to extract the latent features adaptively. Experimental results have shown that our method can work for specific state identification.