One‐Shot Data‐Driven Two‐Degree‐of‐Freedom Controller With Model‐Matching Error Compensation

This study presents an innovative data‐driven design for a two‐degrees‐of‐freedom (TDOF) controller. While many studies, including those employing a data‐driven approach, predominantly concentrate on feedback controllers, the incorporation of feedforward control is crucial for augmenting tracking performance. Traditionally, a feedforward controller uses an inverse system model for the targeted system. However, suboptimal performance may arise if the inverse model demands greater accuracy or external disturbances substantially impact the system. This study introduces a data‐driven TDOF controller that enhances tracking performance and facilitates efficient control system design. Specifically, the TDOF controller is expressed on the foundation of a model‐matching error compensator, with the design integrating a data‐driven approach. Furthermore, sparse regression techniques were applied to alleviate overlearning and compress the controller. Simulation results unequivocally demonstrate that the data‐driven approach yields optimized parameters for the TDOF controller, achieving superior tracking performance compared with conventional methods.