Research on Target Tracking for Robotic Fish Based on Low-Cost Scarce Sensing Information Fusion

Abstract

Target tracking for underwater robots is always challenging, due to low-quality sensing information, information interference, and environmental disturbances. Traditional sensing methods for underwater target tracking include vision-based tracking, acoustic-based tracking, etc., but most of the adopted sensors are extremely expensive or complicated to achieve the mission. In this letter, we investigate the possibility of utilizing low-cost scarce sensing information for underwater target tracking through a robotic fish platform. First, we introduce the design and control system of the robotic fish platform. Instead of using cameras and other expensive sensors, we adopt low-cost infrared sensors as the primary sensors for the robotic fish, which can only detect the appearance of objects within the sensing range of each sensor. Second, we present a target tracking strategy based on the scarce sensing information and fusion by analyzing and combining the information of two adjacent consecutive moments of the sensors. Then, a centralized detection decision tree with fewer branches, fast convergence, and high purity is proposed. Finally, to verify our method, several sets of guided target tracking experiments are conducted. The experimental results show the effectiveness and robustness of the proposed target tracking strategy based on low-cost scarce sensing information fusion.