多变量双线性广义预测极点配置自校正解耦控制器(续)

Generalized Predictive Self-tuning Decoulping Controller with Pole-assingnment for Multivariable Bilinear Systems

对一类多变量双线性系统提出了一种基于预测状态空间实现的GPP自校正控制算法。建立了预测状态与模型结构参数和输入输出信息之间的直接关系。给出了含有多个加权矩阵的多变量二次型性能指标 ,增加了系统设计的自由度。由于加权因子可以根据闭环系统稳定性要求以及系统动态特性、前馈零点增补、输出滤波和跟踪要求分别加以选取 ,可以保证闭环系统稳定并改善了系统动态特性增强了鲁棒性。仿真结果表明了该算法具有GPP的诸多优点。

The paper proposes a Generalized Predictive Pole-assignment self-tuning control recursive algorithm based on predictive state space realization for multivariable Bilinear System. The relationship between predictive state and model construction parameters and input-output information is given directly, so the predictive output matrix of system can be got without using observer. The multivariable object function including multi- weigh matrix for MIMO bilinear system is given, so the free degree of system design is increased. Multi-step predictive output being taken, the information is abundant, and weigh factors may be chosen differently according to the demand of closed-loop system stability as well as of system dynamic feature, feedforward zero-point compensation, output filter and traction, so that the closed-loop system stability is ensured and system dynamic properties are improved and robustness is enhanced. Simulation results show that the algorithm has many advantages of Generalized Predictive Pole-assignment.