为弥补线性系统与非线性系统间的差异, 考虑非线性系统的近似模型-线性参数变化系统的子空间预测控制, 此控制器能更贴切地控制非线性系统.采用数据驱动的子空间预测控制策略, 构造输入-输出观测数据矩阵来辨识状态空间形式下的马尔科夫参数.利用矢量积算子的数据矩阵表示将来时刻的输出预测值, 并以此预测值作用于代价函数.对带有不等式约束的二次代价函数采用并行分布算法来求解其最优值.针对直流电动机在质量分布因素下的线性参数变化系统, 采用子空间预测控制器来控制直流电动机.

To offset the difference between the linear and nonlinear system, here one subspace predictive control for linear parameter varying system which is an approximation of nonlinear system is designed to control the nonlinear system approximately. Applied the data driven method-the subspace predictive control, the Markov parameters coming from the state space form are identified by using the input-output observed data matrices. The future output predictor is denoted by one data matrix and existed in one objective function after introducing some vector product operations. When solving the quadratic cost function with one inequality constraint, the parallel distribution algorithm is used to get its optimum solution. Finally, the proposed subspace predictive control strategy is applied to control the DC motor which exists in one linear parameter varying system with the mass distribution factor.