神经网络监督控制在液压舵机伺服系统中的应用

摘要

Aiming at nonlinear factors in actuator servo system and uncertain interference in working environments, an improved neural network supervisory control was presented to apply in the servo control system design of hydraulic actuator. The controller adopted a single neuron adaptive PID instead of regular liner control to learn, for the purpose of improving robust performance and the stability of system early in the neural networks learning. Then the model of proposed system was founded and verified by Sim-ulink. The result of the simulations indicates that the improved neural network supervisory control can a-chieve an ideal control effect and has strong robustness in the hydraulic actuator servo system, and the proposed strategy carves out a new way for controlling actuator servo system design.%针对液压舵机伺服系统中存在的非线性因素和工作环境的不确定干扰,提出将改进的神经网监督控制算法应用到舵机伺服控制系统设计中.该算法采用单神经元PID控制取代常规线性控制用于神经网络控制器学习,以提高控制系统的鲁棒性及神经网络模型学习初期系统的稳定性.在Simulink环境中建立液压舵机伺服控制系统模型并进行仿真,仿真结果表明:改进的神经网络监督控制,在液压舵机伺服系统中,具有良好的控制效果和较强的鲁棒性,为舵机伺服系统设计提供了一条新的思路.