Deyuan Meng

Robust Design of a Class of Time-Delay Iterative Learning Control Systems With Initial Shifts

This paper is mainly devoted to the iterative learning control (ILC) design for time-delay systems (TDS) in the presence of initial shifts, especially when the system parameters are subject to polytopic-type uncertainties. The ILC laws using a pure error term and/or an initial rectifying action to address the initial shifts are considered, and the two-dimensional (2-D) system theory is employed to develop necessary and sufficient conditions for the asymptotic stability of ILC. For the monotonic convergence of ILC, sufficient conditions are presented in terms of linear matrix inequalities (LMIs) based on the bounded real lemma (BRL). It is shown that adding the pure error term in the D-type learning law helps to meet certain LMIs to achieve a monotonically convergent ILC law. Specifically, this property is first investigated for linear time-invariant systems (LTIS), which is then discussed for the possible extension to TDS. Two numerical examples are included to illustrate the main results.

H ∞ -based design approach to discrete-time learning control systems with iteration-varying disturbances

This paper is concerned with the iterative learning control (ILC) problem for discrete-time systems with iteration-varying disturbances. Using the so-called super-vector approach to ILC, the discrete domain bounded real lemma is employed to develop a sufficient condition ensuring both the stability and the desired H∞ performance of the ILC process. It is shown that this sufficient condition can be presented in terms of linear matrix inequalities (LMIs), which can also determine the learning gain matrix. A numerical simulation example is included to validate the theoretical results.

Frequency-domain approach to robust iterative learning controller design for uncertain time-delay systems

This paper deals with the robust iterative learning control (ILC) for time-delay systems (TDS) with both model and delay uncertainties. An ILC algorithm with anticipation in time is considered, and a frequency-domain approach to its design is presented. It shows that a necessary and sufficient convergence condition can be provided in terms of three design parameters: the lead time, the learning gain, and the performance weighting function. In particular, if the lead time is chosen as just the delay estimate, then the convergence condition is derived independent of the delay and the uncertainties. In this case, with the selection of the performance weighting function, the perfect tracking can be achieved, or the least upper bound of the ℒ2-norm of the limit tracking error can be guaranteed less than the least upper bound of the ℒ2-norm of the initial tracking error.

A particular iterative learning control scheme for linear time-varying continuous systems with input delays

This paper deals with iterative learning control (ILC) design for a class of multi-input multi-output (MIMO), linear time-varying (LTV) continuous systems with input delays. A particular ILC scheme is considered, and is shown to be convergent from the two-dimensional (2D) system point of view. Furthermore, if certain selections of learning gains are met, then this scheme converges in an ordered way, and achieves uniform convergence over the whole learning time interval only after finite iterations. The simulation results show that the proposed scheme can provide different effective updating laws to improve the performance of the control systems.

Monotonically convergent iterative learning control for uncertain time-delay systems: An LMI approach

This paper deals with the robust iterative learning control (ILC) problem for time-delay systems (TDS) subject to matched parameter uncertainties. Based on two-dimensional (2-D) approach, a stability analysis of the ILC process is developed in the sense that the control error converges monotonically as a function of iteration. It shows that a sufficient condition for the ILC stability can be given in terms of linear matrix inequalities (LMIs), which derives learning gains directly. Simulation results show that the ILC system under a law using gains solved by the LMI approach is robustly stable and monotonically convergent.

Gene selection for cancer classification using improved group lasso

An improved group lasso is proposed for simultaneous cancer classification and gene selection. A new criterion is firstly proposed to evaluate the individual gene importance by using the conditional mutual information. Then the weights with biological explanation are constructed and the improved group lasso is presented. A blockwise descent algorithm for solving the proposed model is also developed. The experimental results on lung cancer and prostate cancer data sets demonstrate that the proposed method can effectively perform classification and gene selection.

Solution path algorithm for the fuzzy weighted doubly regularized support vector machine

A fuzzy weighted doubly regularized support vector machine for binary classification is proposed in this paper. Fuzzy weights are presented by using the distance information within each class. Then the fuzzy weighted doubly regularized support vector machine is proposed by combing the weighted hinge loss and the adaptive elastic net penalty. A reasonable correlation between two model parameters is also given and the solution path algorithm to compute the solution paths of the proposed support vector machine is developed. The simulation results on two data sets demonstrate the effectiveness of the proposed method.