论文部分内容阅读
对于一类常见多重时滞非线性离散系统．提出了基于动态线性逼近的增量型最小化模型、递推预测 模型，无模型学习自适应控制律和带有参数限定时域长度的参数自适应预报递推算法，实现了对存在较大滞 后的多重时滞非线性系统的无模型学习自适应控制。该算法不需要受控系统的结构信息、数学模型、外部实 验信号和训练过程，不用解Ｄｉｏｐｈａｎｔｉｎｅ方程．无需矩阵运算．在线计算量很小．实时性好，仅用受控系统的 Ｉ／Ｏ数据来设计，传统的未建模动态不存在。通过仿真表明，该算法对于一类非线性系统实现无模型自适应 控制是正确和有效的．
For a class of common nonlinear time-delay discrete systems. An incremental minimization model based on dynamic linear approximation, a recursive prediction model, a model-free adaptive learning control law and an adaptive parameter prediction recursive algorithm with parameter-limited time-domain length are proposed, Modelless learning adaptive control for multiple time-delay nonlinear systems. The algorithm does not need the structural information of the controlled system, the mathematical model, the external experimental signal and the training process, without solving the Diophantine equation. No matrix calculation. Online calculation is small. Good real-time, only the controlled system I / O data to design, the traditional unmodel dynamic does not exist. Simulation results show that the proposed algorithm is correct and effective for a class of nonlinear systems to realize model-free adaptive control.