The Low Cost DOA estimation Method for LSS-Target via Intelligent Reflecting Surface System and Deep Neural Network

Detection and localization LSS (Low altitude, Slow speed and Small) targets, having become a cosmopolitan difficult problem and studying hotspot. To address this challenge, a direction of arrival (DOA) estimation method in intelligent reflecting surface(IRS) with low-cost ADC is first investigated in this paper. We propose a deep neural network (DNN) as recovery method for the low-resolution sampled signal, which can estimates the quantization noise caused by the ADC accurately. Then, the denoised signal is utilized to improve the DOA estimation, and the recovery performance for the quantized signal is evaluated by DOA estimation. Simulation results show that, the better reconstruction performance is achieved by the proposed network than existing denoising method under the same training conditions. Through comparison with state-of-the-art methods, DNN shows a better denoising effect. Combined with DNN denoising, the performance of the DOA estimation with 1-bit ADC outperforms existing 2-bit ADC systems.