基于同伦BP网络的参数自调整ANN-PI控制器

Parameter Self-adjusting ANN-PI Controller Based on Homotopy BP Networks

BP神经网络可有效地记忆模糊控制规则,并以"联想记忆"方式使用这些经验。然而,现有前向网络学习算法不可避免地存在局部极小问题,同伦连续BP算法可有效地解决BP网络的全局收敛性问题,同时使网络具有很快的收敛速度。为了进一步提高控制系统的精度和抗干扰能力,设计了一种参数自调整ANN-PI控制器。实验结果表明,这种控制器动态响应快,控制精度高,抗干扰能力强,对参数变化不敏感,具有一定的鲁棒性。

Back-Propagation neural network can record the fuzzy control rules efficiently,and utilize these experiences according to associative memory.All now unavailably feedforward net learning algorithms have a local minimum problem.The homotopy continuation BP algorithm provides an effective method for global convergence of BP network and is of very fast convergent speed.To enhance control system’s accuracy and disturbance-resisting ability,an ANN-PI controller with the parameters self-adjusting was designed,which has a faster dynamic response,higher control accuracy,better disturbance-resisting ability,less sensitive to parameter changes,and robustness.