Modeling and Control of a Bioinspired Robotic Fish Underwater Vehicle

Robotic fish is propelled by an oscillating, 2-joint caudal fin and two pectoral fins. Major kinematic parameters of robot are based on kinematics study of a yellow fin tuna. A bio-inspired algorithmic framework with these parameters is found to produce the body wave, along robot's joints. A novel dynamic model is proposed by unifying conventional rigid body dynamics and bio-fluid-dynamics of carangiform swimming. Accuracies of hydrodynamic forces are examined. A central pattern generator structure is applied to generate desired rhythmic patterns preserving control properties. A two level locomotion control architecture based on vertebrate fish biology is implemented. An inverse dynamic control method based on non-linear state function model is proposed to improve tracking performance. Further, dynamic motion closed loop control strategies are developed, implemented and compared based on three different nonlinear control schemes. Experimental results obtained show that inverse dynamic model based control using dynamic compensation can improve trajectory tracking accuracy and smoothness significantly. Robotic fish is found to exhibit swimming patterns similar to the biological fish.