Target Tracking Strategy of Robotic Fish Based on Multi-Sensor Distributed Detection Information Fusion

Abstract

Target tracking for robotic fish is always challenging due to complex fluid dynamics and inherent yaw instability. This paper introduces the design of the robotic fish and presents the target tracking strategy of the robotic fish based on multi-sensor distributed detection information fusion. The robotic fish adopts a semi-sealed design, which is composed of three parts: a rigid head, a wired-driven active body, and a compliant tail. An improved CPG model is utilized for robotic fish to imitate the rhythmic movement of real fish. Instead of using cameras and other expensive sensors, we adopt a low-cost infrared sensor with scarce sensing information as the primary sensor. And then, the target tracking strategy is based on multi-sensor distributed detection information fusion by analyzing and combining the information of two adjacent consecutive moments of the sensors. Finally, several sets of experiments are conducted to verify the effectiveness and robustness of the target tracking strategy.