Learning and Operator Based Control System Design for Soft Robotic Finger with Denial of Service Attack and Input Hysteresis Using Right Coprime Factorization

摘要

Network attacks generally occurred in practical control system and possibly cause instability even insecurity. Specifically, Denial of Service (DoS) attack is the one that could cut down the connection between equipment and make the feedback signal zero consequently. In this paper, a machine learning based control system design is proposed to deal with the problems caused by DoS attack and input hysteresis in the case of soft robotic finger. The system is accomplished using right coprime factorization (RCF) theory, by which the robust stability of proposed system is guaranteed. In addition, the results of practical experiments are exhibited as well to verify the effectiveness of proposed system.