A novel bionic mantis shrimp robot for tracking underwater moving objects

Abstract

ABSTRACT Inspired by the agility of mantis shrimp, this paper presents the development of a novel bionic mantis shrimp robot with multiple pleopods coupled motion for underwater target recognition and tracking. Firstly, the novel bionic robot is constructed based on the inspiration from biological mantis shrimp. Secondly, a Kalman filter-based algorithm, namely MobileNet-YOLO (KMY), is created to collect datasets and train neural networks. Thirdly, a moving target following system is developed for the bionic mantis shrimp robot, in which a dual PID controller is used to track underwater moving targets. The real robot testing results show that the proposed control system can enable our bionic robot to follow a specific moving target in a narrow pool (2m × 1m × 1m), and the minimum turning radius can be up to 0.55m when the angle between the robot's initial motion and the motion of target is 90°.