Visual Tracking of Multiple Robotic Fish for Cooperative Control

In this paper, we investigate visual tracking of multiple biomimetic robotic fish which swim together to achieve some specific tasks in underwater cluttered environments. Taking account of surrounding background and kinematic characteristics of swimming fish, a novel color-index-based identification approach is presented, which is capable of identifying many fish rapidly in a processing cycle. Meanwhile, some anti-jamming measures including optical correction and foil superposition are made to ensure a robust tracking. Finally, all tracking operations are optimized in coding with the aid of parallelized SIMD technologies embedded in processors. Our proposed visual system can experimentally track as large as an amount of eight robotic fish and one obstacle within 24.143 milliseconds, which can fully satisfy the requirements of cooperative control.