Altitude measurement based on characteristics reversal by deep neural network for VHF radar

A novel direction of arrival (DOA) estimation method is proposed for very high‐frequency (VHF) radar by the deep neural network (DNN) under strong multipath effect and complex terrain environment. The classical methods are all based on the classical multipath signal model, hence, it often causes the problem of model mismatch and results in poor performance in estimation. It is generally considered that the serious multipath effect reduces the precision of elevation estimation. However, the characteristics of the multipath signal are exploited and used to improve the precision in this study. This is the highlight of the proposed method. The approach of the deep neural network is applied to learn the received data's characteristics from a different elevation. A new characteristic space is constructed in the training procedure. In the test procedure, the characteristic of data is extracted by the well‐trained network and projected into the constructed characteristic space. Reversing the DOA is finished at last. The results of simulation data verify the validity of the proposed method. The results of the practical data show the practicability of the proposed method for the low‐elevation target under the serious multipath effect. Combining the DNN approach with altitude measurement of a low‐elevation target for VHF radar is meaningful to explore.