Active Fault Detection of Soft Manipulator in Visual Servoing

This article proposes a novel method as an improvement of the active fault detection method for eliminating the negative influences generated by accessorial signals used in it. In this method, the signals used for fault detection are generated by the controller and are merged with control signals and, therefore, the stability of the system can be rigorously proved. A specific example of the method is presented to detect the locked-motor faults and finish the visual servoing task of a soft manipulator. In the example, the controller stabilizes the system and generates signals used to detect locked-motor faults. In order to enhance the robustness of the method, an artificial potential field is introduced to impose constraints on the Jacobian matrix of the soft manipulator. Based on it, the adaptive algorithm is designed to guarantee the stability of the system, and the convergence of the image error is proved. Simulations and experiments are conducted to verify the performance of the proposed controller and results demonstrate that on the one hand, the system is stable and the image errors converge to zero when no faults occur; on the other hand, the locked-motor faults can be detected quickly and precisely when it happens.