神经网络控制、无模型控制PID控制仿真比较

Simulation Comparisons among Neural Networks Control, Model-free Adaptive Control and PID Control

在同一条件下,对三种仅用受控系统 I/O 数据来设计控制器的方法进行了控制性能仿真比较,三个受控系统分别取自文[1]和[5]。三种不同的控制方法各有其优缺点,仿真结果表明:无模型自适应控制具有其他两种方法所不具备的优点,适合处理结构时变、参数时变、阶数时变非线性离散时间系统的控制问题。

Under the identical conditions, three different control strategies only using I/O data are compared through simulations. Simulation results indicate that the model-free control method is superior to the other two methods because it can deal well with the control problem of the discrete time nonlinear plant whose structure, parameters and order are time-varying.