Application of BP neural network to the data classification for power testing of the microwave circuit board
Long Yuan, Zhao Xiucai, Guo Rongbin · 2015
The microwave circuit power test board is designed to automatically gather voltages and powers based on the multi-channel analog signal sensor. How to evaluate the rationality of the gathered voltages and powers is an important research topic. In this paper, a classification algorithm based on BP neural network is applied to estimate voltages or powers in each channel. A modified gradient descent method with the additional momentum and variable learning rate is proposed to optimize the descent direction. Finally, numerical results are presented toverify that the proposed method has high accuracy.