Collision-free motion planning for a biomimetic robotic fish based on numerical flow field

In this paper, we investigate motion planning for a biomimetic robotic fish which swims in underwater cluttered environment. For the collision-free motion planning problem, a widespread approach is to decouple it into two subproblems: paths generation and tracking control (or feedback regulation). However, consider the special propulsion mechanism and kinematic property of the robotic fish, it is difficult for it to track a predetermined path. To cope with this situation, a novel approach based on numerical flow field is proposed, which will generate a feasible velocity vector field instead of a determined path. Then the robotic fish will be inducted by a series of feasible velocity vectors to reach its destination. Experiments are conducted to verify the effectiveness of our approach