Cooperative Control of Multiple Biomimetic Robotic Fish
Junzhi Yu, Min Han Tan, Long Wang · 2008
This chapter has reviewed some of the issues involved in creating a multiple robotic fish cooperation platform inspired by the astonishing cooperative power exhibited by fish school. Grounded on an optimized kinematic and dynamic model of robotic fish, a group of radio-controlled, multi-link fish-like robots as well as their motion control were developed. To enable a closed control loop, a vision-based multi-object tracking subsystem for multiple robotic fish was built, where an improved parallel visual tracking method is formulated within a framework synthesizing features of fish features and surrounding disturbance. A hierarchical architecture for the artificial multi-fish system was further proposed to generate coordinated behaviours and actions. Experimental results on three typical cooperation tasks demonstrated the effectiveness of the proposed scheme. The ongoing and future work will focus on developing a 3-D dynamic model for freeswimming, multi-link robots and on investigating experimentally the corresponding control laws, which makes preparations for cooperative control of multiple robotic fish in 3-D aquatic environments. In addition, some learning algorithms will be incorporated into the cooperative framework to enhance the cooperation efficiency.