Two-Degree-of-Freedom Controller Design Based on a Data-Driven Estimation Approach

Abstract

Data-driven control design is a method to create and tune controllers directly from the initial experimental data without a mathematical model to be controlled. Tracking and disturbance suppression are necessary to control real systems. A two-degree-of-freedom (2DOF) control system is effective to simultaneously enhance the performances of both. This study proposes a direct data-driven tuning method for the controller parameters of a 2DOF control system using only one-shot initial experimental data without mathematical modeling of the controlled object. The proposed approach improves the tracking and disturbance suppression performances by utilizing an estimation method in which the sensitivity function and the closed-loop transfer function are identified after updating the parameters in the time domain. Specifically, the closed-loop response and control input are estimated after updating the parameters, realizing efficient control system design because the control performance can be evaluated prior to implementation. To validate the effectiveness of the proposed method, a simulation for a mechanical system and an experiment for motor control are performed. The proposed method can estimate the response and control input of a 2DOF control system after updating the parameters by offline computations. The optimal control parameters can be obtained to enhance the tracking and disturbance suppression performances.

Direct tuning of the data-driven controller considering closed-loop stability based on a fictitious reference signal

“…In general, experimental data is corrupted by measurement noise. Although the noise may affect the performance of data-driven control, many studies have examined strategies for addressing noise, such as regularization [36], total variation denoising [53], discrete Fourier transform via periodization [54] , and reformulation of the optimization problem using slack variables [57]. In the future, the noise tolerance of the proposed method will be evaluated and an appropriate handling method for noise will be explored.…”

“…However, if perfect model matching is not achievable by the controller class considered in design, the closed-loop system given by the tuning result may be unstable, as discussed in [53]. Several authors have proposed a modified version of FRIT, i.e., instability-detecting FRIT (IDFRIT), so as to avoid instability [53,54].…”

A Physics-Guided Data-Driven Feedforward Tracking Controller for Systems With Unmodeled Dynamics—Applied to 3D Printing

“…Moreover, when the stability or convergence is being analyzed, the broad knowledge of plant dynamics is typically required [16]. Adaptive FF controllers, on the other hand, tune the parameters of the controllers to account for unmodeled or changing dynamics [18], [19], [20], [21], [22], [23]. While there are some adaptive FF controllers that must be tuned iteratively using repeating trajectories or trials, e.g., [18], [20], and [21], literature has been exploring the approach to improve the extrapolation performance of the iterative tuning process [18], and besides not all adaptive FF controllers need this requirement.…”

“…While there are some adaptive FF controllers that must be tuned iteratively using repeating trajectories or trials, e.g., [18], [20], and [21], literature has been exploring the approach to improve the extrapolation performance of the iterative tuning process [18], and besides not all adaptive FF controllers need this requirement. However, a major shortcoming of adaptive FF controllers is that they typically require a fixed structure with associated parameters to be tuned online [18], [19], [20], [21], [22], [23], despite the fact that the stability properties can be obtained more easily due to fixed structure. This diminishes their usefulness in situations where the exact structure of the unmodeled dynamics is a priori unknown or changing, such as the unknown aftermarket modification of 3D printers.…”