【摘要】 提出基于变神经网络学习动态系统参数的最小方差预测控制器。其目的是通过在线学习 ,使控制器(MVPC)能适应被控对象参数变化和非确定性。提出的变神经网络由两部分组成 ,一部分是线性神经网络 (LNN) ,作为被控对象局部线性动态模型 ,另一部分是多层交叉回归神经网络 (DRNN) ,它近似为非线性动态模型。由于引进递推最小方差算法 ,本控制器运算速度相当快。仿真结果表明所提方法对非线性系统自适应控制是有效的更多还原

【Abstract】 This paper presents a minimum variance predictive controller(MVPC) basing a modified Neural network(MNN) in order to learn the characteristics of a dynamic system. The purpose is the MVPC can adapt parameters’ variation and uncertainty in the controlled plant through the on-line learning. The paper presents network is composed of two parts: one is linear neural network(LNN), which models the local linearisation dynamics of the controlled plant, and the other is multilayered diagonal recurrent neural network(DRNN) which approximates the nonlinear dynamics not being modeled by the linear model. The learning algorithm is considerably faster because of the introduction of recursive least squares(RLS) algorithm. Simulation results have shown that the proposed approach is effective for adaptive control of nonlinear systems.更多还原