关键词: 机电作动器, 永磁同步电机, 对角递归神经网络, 自适应学习率

Abstract: Aiming at the stability and response speed requirements of electromechanical actuator (EMA), a novel speed loop control method of EMA based on diagonal recursive neural network (DRNN) was proposed. The gain parameters were dynamically adjusted in real time. Aiming at the problem that fixed learning rate cannot strike a balance between the system stability and rapidity , an adaptive learning rate algorithm was designed to enhance the learning ability of the controller. The simulation results showed that, the improved DRNN-PI control method improved the dynamic response ability and anti-interference ability of EMA speed loop, and could effectively ensure the stability and robustness of the overall EMA system compared with the traditional PI control, fuzzy-PI and the fixed learning rate method.

Key words: electromechanical actuator(EMA), permanent magnet synchronous motor, diagonal recursive neural network(DRNN), adaptive learning rate