神经网络辨识的自适应逆控制

逆模型控制是一个新颖的控制方法．但在实现上会遇到很多困难，如被控对象的大滞后、时变性和不确定性等，使精确的对象数学模型难以建立．文中根据工业对象的特点及对控制系统高鲁棒性与高自适应性的要求，提出一种改进的神经网络的模型参考自适应逆控制系统．仿真试验表明，此系统具有良好的跟踪给定信号和消除对象干扰的作用．

Adaptive Inverse Control Based on Neural Network Identification

Inverse model control is a novel control method, However, it will meet a lot of difficulties in realization, such as large time delay of controlled plant, time varying and uncertainty, so that it could be hard to establish accurate mathematical model. According to the character of industrial plants and the need for high robustness and high adaptability of control systems, the author gives here a model reference adaptive inverse control system based on the improved neural network. As shown in simulation test, this system possesses good effect of tracing preset signal and cancelling interference on the plant.