Bioinspired Closed-loop CPG-based Control of a Robotic Manta for Autonomous Swimming
Yiwei Hao,Yonghui Cao,Yingzhuo Cao,Xiong Mo,Qiaogao Huang,Lei Gong,Guang Pan,Yong Cao
DOI: https://doi.org/10.1007/s42235-023-00424-z
IF: 2.995
2024-01-01
Journal of Bionic Engineering
Abstract:Fish in nature exhibit a variety of swimming modes such as forward swimming, backward swimming, turning, pitching, etc., enabling them to swim in complex scenes such as coral reefs. It is still difficult for a robotic fish to swim autonomously in a confined area as a real fish. Here, we develop an untethered robotic manta as an experimental platform, which consists of two flexible pectoral fins and a tail fin, with three infrared sensors installed on the front, left, and right sides of the head to sense the surrounding obstacles. To generate multiple swimming modes of the robotic manta and online switching of different modes, we design a closed-loop Central Pattern Generator (CPG) controller based on distance information and use a combination of phase difference and amplitude of the CPG model to achieve stable and rapid adjustment of yaw angle. To verify the autonomous swimming ability of the robotic manta in complex scenes, we design an experimental scenario with a concave obstacle. The experimental results show that the robotic manta can achieve forward swimming, backward swimming, in situ turning within the concave obstacle, and finally exit from the area safely while relying on the perception of external obstacles, which can provide insight into the autonomous exploration of complex scenes by the biomimetic robotic fish. Finally, the swimming ability of the robotic manta is verified by field tests.