Complex-Value Convolutional Neural Network for Classification of Human Activities

Human target is one of the main targets of radar detection. Many methods have been proposed to improve the performance of human activities classification, including the neural networks methods. In the current research, the neural network methods only utilize the amplitude information of the target echoes and ignores the phase information. To solve this problem, this paper proposes a human activities classification method based on Complex-Value Conventional Neural Network (CV-CNN). In this method, the motion capture data are utilized to obtain the motion trajectories of various parts of the human, the time-frequency (TF) analysis is utilized to obtain the TF representations, and a CV-CNN structure is built to train the complex-value TF data of different human activities. The experiment results verify the effectiveness of the proposed method. Meantime, the results under different SNRs demonstrate the robustness of the proposed method.