Autonomous swimming on limit cycles with disturbance rejection capability for a fish-inspired robot.

Position-predefined velocity control in the presence of parametric uncertainties as well as the ability to reject external disturbances simultaneously, has been one of the challenges in the area of fish-like swimming robots. To overcome the abovementioned problems, in this article, after identifying parameters of the Chapylgin sleigh model (CSM), which is a low-dimensional time-varying nonholonomic system with limit cycle dynamics, a chattering-free robust backstepping controller (RBSC) and a disturbance observer (DO) are proposed for autonomous position control with external disturbance rejection capability in the presence of uncertainties in velocity space for an under-actuated swimming robot with an internal reaction wheel. The uniform boundedness of the whole system, including RBSC and DO, is theoretically performed by Lyapunov methods. Because of the inherent limit cycle dynamics in the reduced velocity space of the swimming robot, pure usage of the observer results in fluctuating estimations of disturbances. Thus, the moving average technique is employed to enhance the accuracy of the estimated disturbance signals. The simulations are performed for various trajectory tracking tasks, and the results demonstrate the high maneuverability and robustness of the proposed schemes in the perturbed conditions. In addition, computational fluid dynamic validation simulations confirm the reliability of the proposed CSM-based feedback controller in the presence of high-dimensional fish-fluid interaction with the vortex wake of a hydrofoil. [ABSTRACT FROM AUTHOR]