Radar High Range Resolution Rrofile Attitude Estimation Based on Neural Network

The sensitivity of attitude is core problem of radar high range resolution profile (HRRP) target recognition, if the orientation of targets can be approximately confirmed, we can narrow the search area of template library in recognition. Then the real time of recognition results can be guaranteed, and the recognition performance can be improved too. In this paper, we propose a radar high range resolution profile attitude estimation method based on neural network, the data argumentation and regularization technique improve the generalization ability of model, and they make the model have some noise robustness.