Nonlinear observer design for stabilization of gliding robotic fish

Gliding robotic fish is a new type of underwater robots for aquatic environmental monitoring applications. The robot features both buoyancy-driven gliding and active tail-actuated swimming or maneuvering. In our previous work, a passivity-based controller was designed for stabilizing the glide motion. Estimation of the system states including the velocities, which is essential for the feedback control implementation, is challenging for such a low-speed and low-cost underwater robot with highly nonlinear dynamics. In this paper, a nonlinear model-based observer is proposed to estimate the system states. For comparison purposes, a linear observer based on linearized dynamics is also considered. The nonlinear observer proves to have better performance over the linear observer, especially in terms of the robustness against the measurement noise. The proposed nonlinear observer is then implemented in the passivity-based stabilization controller on a gliding robotic fish, and the experimental results are presented to support the effectiveness of both the controller and the observer.