.png){kind=icon}
.png){kind=icon}
Return
Factorized VRFT: Proposal and Experimental Verification
.png){kind=icon}
OA
AI
DOI：10.1109/ACCESS.2026.3657809.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This paper discusses the data-driven model matching problem with a one-shot preliminary experiment. Although many methods, such as Virtual Reference Feedback Tuning (VRFT), have been proposed as a solution for this problem, they are nonconvex in general, i.e., convex only under the assumption on the structure of the feedback controller. This paper introduces a new method, named Factorized-VRFT (F-VRFT), as a solution for the above problem, where the cost function is convex for any linear controller. In particular, the cost function of F-VRFT is quadratic; thus, the optimal parameter is available explicitly and the computational time becomes less than 1/100 of that of VRFT. Furthermore, by leveraging this property, we show that the method can be easily extended to adaptive model matching. The effectiveness of this method is demonstrated through experiments with a practical motor.
Keywords:
Data-driven control
model matching
RLS
VRFT
Journal
IF:
3.6
Papers:
9.8W
Citations:
29.4W