Atomic Network-Based DOA Estimation Using Low-Bit ADC

In the direction of arrival (DOA) estimation problem, when a low-bit analog to digital converter (ADC) is used, performance severely deteriorates. this paper, DOA problem considered in low-cost finding system with ADC. To eliminate quantization noise, we propose novel network ADCnet, which composition fully connected layers and exponential linear unit (ELU) layers, input signals are received using After an AtomicNet also proposed estimate from denoised signals, where atomic vectors corresponding steer vectors. A loss function considering both reconstruction sparsity AtomicNet. Different exiting norm-based methods, method can avoid optimization lower computational complexity. Simulation results show that outperforms existing methods