Target tracking of robotic fish based on embedded vision and CPG model

Abstract

There has been greatly increased interest in using embedded vision for aquatic environments. This paper presents a robust target tracking method for a free-swimming biomimetic robotic fish by means of embedded vision and central pattern generator (CPG)-based motion control. Specifically, an algorithm called automatic continuous adaptive mean shift (Auto-CAMSHIFT) is firstly proposed to obtain the position and size of the interested target. Then a fuzzy logic controller is developed to generate the control input closely related to the desired target. A CPG controller that is robust against small and unexpected disturbance, at the same time, is employed to produce coordinated signals for multiple moving joints of the robotic fish. Finally, aquatic testing results verify the effectiveness of the proposed methods and show a satisfactory performance.