A DMD-Based Directional Stability Control of an Anguilliform Swimming Continuum Robot

Biorobotics employs principles of natural locomotion to improve robots' mobility. Among various swimming modes, anguilliform locomotion is particularly recognized as an energy-efficient mode that incorporates complex physics. Due to whole-body undulation, the determination of the anguilliform swimmer’s direction is not trivial, and the mechanism of muscle interaction to maintain the desired direction is complex. In this study, the problem of predicting and controlling the gross motion trajectory of a soft robot with anguilliform swimming on the water surface is investigated. The robot consists of a six-segment continuous body, where each segment is actuated with artificial muscles to bend left and right. A DMD-based mode extraction technique is proposed to identify the robot's trajectory, which represents the heading direction. The algorithm, namely CDE DMD, utilizes delay-embedded complex variable states to calculate the future trend of the robot state. For the control, a hypothesis that asymmetric sidewise actuation results in regulation of the swimming direction was investigated using COMSOL Multiphysics® 6.2. The results demonstrate the CDE DMD ability to predict gross motion for various scenarios. The prediction and asymmetric actuation rules can be used to maintain swimming in a straight line, a problem which is known as directional stability.