为解决预测控制综合方法中的模型不确定问题,不同于以往利用多胞模型描述的方法,提出了一种新的基于递推子空间的自适应预测控制综合方法.通过在每一步中加入当前输入输出数据重新构建Hankel矩阵,对广义能观矩阵进行更新,从而获得对应的状态空间模型;然后将新获得的模型应用于预测综合的优化求解过程,得到当前时刻的控制律.为提高算法的收敛速度,在辨识的过程中引入了基于模型匹配误差的时变遗忘因子.最后,在慢时变与线性时不变两种情况下进行仿真,验证了所提出算法的有效性.
To address the model uncertainty problem of the synthesis approach of the predictive control, a new synthesis method of adaptive predictive control based on recursive subspace which is different from the previous methods with polytopic description is proposed. Hankel matrix is rebuilt by adding the current input and output data in each step, and then the extended observability matrix is updated. The corresponding state space model is obtained and then is used in current control of optimization solution to get the current control law. To improve the convergence rate of the algorithm, a model matching error based time-varying forgetting factor is introduced into the identification process. Finally, two simulation examples are presented in the linear time-invariant and the slow time varying situations, and the results verify the effectiveness of the proposed algorithm.