基于ANFIS的迭代学习控制初始控制研究

Study of Initial Iterative Learning Control Strategy Based on Adaptive Neural-Fuzzy Inference

将自适应神经模糊推理算法用于迭代学习控制初始控制输入量的求取问题,提出一种基于自适应神经模糊推理系统的迭代学习初始控制算法。针对传统迭代学习控制中对于经验数据运用不足或是没有运用的问题,利用迭代学习控制对于以往控制任务的学习建立经验数据库,通过自适应神经模糊推理系统对于经验数据库中的数据进行拟和以得到新的控制输入量。通过仿真分析验证了算法的可行性和有效性。

In order to avoid the blend choice of the initial control input in iterative learning control when the control system faced a new desired trajectory-tracked task or a new environment,an improved algorithm is proposed to obtain the initial value of the iterative learning control based on adaptive neural-fuzzy inference.Desired control input of iterative learning control is estimated by adaptive neural-fuzzy inference incorporated experience database.The simulation results show that the convergence speed is satisfactory regardless.