中国科学院自动化研究所机构知识库

A bio-inspired marsupial robotic fish system composed of heterogeneous robotic fish is proposed in this paper. A miniature robotic fish, as the daughter robotic fish, can adapt to some narrow space, while the mother robotic fish, with a cabin to transport the daughter, possesses a powerful movement ability to improve the mobility and endurance of the team. The structures for mimicking bio-motion and the method for fishlike-motion are presented. A typical task of daughter-mother following is given to show the cooperation of the team. A motion model of free swimming is built based on the Lagrangian function, and the coupled dynamic and kinematic functions are calculated based on the relation between the generalized force and fluid force. Then, a neural network is trained through the data generated from this model to get a predictive yaw controller, which can control the orientation by a different offset of each link. The daughter robotic fish adopts a dynamic light source tracking approach to follow the mother, and a heterogeneous communication-based finite state machine is presented for task modeling. Experiments are carried out to verify the system.