DOA Estimation Based on Neural Network for HFGWR
基于神经网络的高频地波雷达目标到达角估计

Direction of Arrival (DOA) technology based on neural network is discussed. In the paper, the DOA method based on function approach and model classification with coding is presented and it employs two kinds of neural networks: Radial Basis Function Network(RBFN) and General Regression Neural Net (GRNN). The paper introduces the network structure, simulation and the application to High Frequency Ground Wave Radar(HFGWR). The simulation and real data processing verifies that the model classification method based on GRNN offers better performance than others, its performance is good even the signal-to-noise of the signal is low until 4 dB.