Abstract:Predictive control, which is based on a model of the system to compute the applied input optimizing the future system behavior, is by now widely used. If the nominal models are not given or are very uncertain, data-driven model predictive control approaches can be employed, where the system model or input is directly obtained from past measured trajectories. Using a data informativity framework and Finsler's lemma, we propose a data-driven robust linear matrix inequality-based model predictive control scheme that considers input and state constraints. Using these data, we formulate the problem as a semi-definite optimization problem, whose solution provides the matrix gain for the linear feedback, while the decisive variables are independent of the length of the measurement data. The designed controller stabilizes the closed-loop system asymptotically and guarantees constraint satisfaction. Numerical examples are conducted to illustrate the method.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to use data - driven methods to design robust model predictive control (MPC) schemes in the presence of uncertainties, so as to handle the input and state constraints of linear parameter - varying systems (LPV) and Lur'e - type nonlinear systems. Specifically, the paper proposes a data - driven robust MPC method based on linear matrix inequalities (LMI), which can directly obtain the system model or input through historical measurement trajectories without an accurate system model, and ensure the asymptotic stability of the closed - loop system and the satisfaction of constraints. ### Main contributions: 1. **Data - driven framework**: The paper proposes a framework based on data informativeness and uses Finsler's lemma to design the controller. The advantage of this method is that the decision variables are independent of the time length of the experimental data, which is suitable for large - scale data sets. 2. **Robustness**: The paper considers two types of uncertain systems: slowly varying linear systems and Lur'e systems with uncertain nonlinearities. For these two systems, the paper respectively establishes optimization problems in the form of LMI to ensure that the controller can still stabilize the system and meet the constraint conditions in the presence of uncertainties. 3. **Theoretical guarantee**: Through the optimization problems in the form of LMI, the paper proves that the designed controller can asymptotically stabilize the closed - loop system and ensure the satisfaction of input and state constraints. 4. **Numerical verification**: The paper verifies the effectiveness of the proposed method through two numerical examples. The first example is a linear time - varying system, and the second example is a flexible manipulator system with sector - bounded nonlinearities. ### Technical details: - **Data informativeness**: The paper defines the concept of data informativeness, that is, whether a given data set is sufficient for designing a controller that meets specific performance requirements. - **Finsler's lemma**: Use Finsler's lemma to transform the data - driven control problem into an optimization problem in the form of LMI, so that it can be solved using existing optimization toolboxes. - **Optimization problems**: The paper proposes LMI optimization problems for nominal systems, linear parameter - varying systems, and Lur'e - type systems. These problems are in the following forms: - **Nominal system**: \[ \begin{aligned} & \min_{N, L, \alpha, \eta, \epsilon > 0} \quad \alpha \\ & \text{s.t.} \quad \begin{bmatrix} 1 & x_0^\top \\ x_0 & N \end{bmatrix} \succ 0, \\ & \quad \begin{bmatrix} N - \eta I & 0 & 0 & 0 \\ 0 & 0 & 0 & N \\ 0 & 0 & 0 & L \\ 0 & N & L^\top & \Psi^\top \\ 0 & 0 & 0 & \Psi & \alpha I \end{bmatrix} + \epsilon \begin{bmatrix} X^+ \\ -X^- \\ -U^- \\ 0 \\ 0 \end{bmatrix} \begin{bmatrix} X^+ \\ -X^- \\ -U^- \\ 0 \\ 0 \end{bmatrix}^\top \succ 0, \\ & \quad \begin{bmatrix} N & \Psi^\top \\ \Psi & \alpha I \end{bmatrix} \succ 0, \\ & \quad \begin{bmatrix} 1 & d_i L + c_i N \\ (d_i L + c_i N)^\top & N \end{bmatrix} \succ 0 \end{aligned} \]

LMI-based Data-Driven Robust Model Predictive Control

Matrix Inequalities Based Robust Model Predictive Control for Vehicle Considering Model Uncertainties, External Disturbances, and Time-Varying Delay

Robust Predictive Control of Uncertain Intergrating Linear Systems with Input Constraints

LMI Based Multi-Model Robust Predictive Control

Event-Driven Robust Output Feedback Control for Constrained Linear Systems Via Model Predictive Control Method

Robust Data-Driven Predictive Control for Unknown Linear Time-Invariant Systems

Design and Implementation of Novel LMI-Based Iterative Learning Robust Nonlinear Controller

Constrained Robust Model Predictive Control Embedded with a New Data-Driven Technique

Provably Safe and Robust Learning-Based Model Predictive Control

Robust Model Predictive Control For Nonlinear Systems

Robust Output-Lifted Learning Model Predictive Control

Improving Robust Model Predictive Control Via Dynamic Output Feedback

Robust Learning Model Predictive Control for Linear Systems Performing Iterative Tasks

Robust Parameter-Dependent Constrained Model Predictive Control

Linear programming-based robust model predictive control for positive systems

A Robust Model Predictive Control Scheme for Constrained Linear Systems

Robust predictive control for polytopic time singular system with input constraints

Data-Driven Min-Max MPC for Linear Systems: Robustness and Adaptation

Optimal Data-Driven Prediction and Predictive Control using Signal Matrix Models

An LMI approach to robust model predictive sampled-data control for linear uncertain systems

Data-Driven Model Predictive Control With Stability and Robustness Guarantees