直接作用式神经网络控制器研究

A New Direct Neural Network Controller and Its Application in Milling

提出了一种新颖的直接作用式神经网络控制器，将其应用于铣床切削过程控制，使铣削力在加工条件变化时始终保持恒定．该控制器具有在线自学习能力，无需预先训练网络的权值，更不用建立控制对象的数学模型．实验结果表明其控制性能优越，自适应能力强．

A new direct neural network controller is developed, which has been successfully used in the milling process control and a constant milling force under variable milling conditions is maintained. The controller is capable of on-line self-learning and neither the pretraining of network weights nor the development of a mathematical model for the control object is required. Simulation and experimental results show that it is of excellent performance and with good adaptive ability.