论文简介 |
In this article, a reinforced gradient-type iterative learning control algorithm is developed for a type of
discrete linear time-invariant systems with parameters uncertainties and external noises. The technique is to
construct a symmetric learning gain matrix on basis of the system Markov parameters and an appropriate
learning step length. First, for the case when both the model uncertainties and the external noises are
absent, sufficient and necessary monotone convergences of the proposed algorithm are derived by means
of matrix theory and norm inequality under the assumption that the learning step length is properly chosen.
Then, for the cases when the model uncertainties are tolerable and the external noises are bounded, the
robust monotone convergence and robustness are respectively analysed. Compared with the conventional
gradient-type iterative learning control scheme, the proposed reinforced one is more efficient in speeding
up the convergent tracking performance and resisting perturbations. Numerical simulations testify the
validity and the effectiveness as well as the feasibility. |