One-Bit DOA Estimation Based on Deep Neural Network

This paper established a deep neural network model for DOA estimation of narrowband signals. First, one-bit quantization is considered into implementation for only retaining the symbol information of training data, as it offers low cost and low complexity in actual communication system. Then we investigate the performance of the neural network trained with quantized data and traditional MUSIC algorithm. Finally, simulations are conducted for correctness and validation. The results illustrate that the proposed method can realize meshless DOA estimation and has higher estimation accuracy in the case of low signal-to-noise ratio.