Formation Control Of Multiple Robotic Fish

With the development of robotics investment, land-based robots cannot meet the demand for human exploration to the environment and improvement of living. Robots are in urgent need in all areas, including oceanic supervision, military detection seabed exploration etc. Most of developed countries have established specialized research institutes for underwater robots. But many researches focus on the underwater robot model and mechanical structure, only few of them is concentrated in task analysis and intelligent control.In the fields of robotics research, it is aware gradually that the achievements of multi-robot cooperative are far more than the single robot do. The formation control has been an important research topic recently. There have been a lot of perfect control strategies, which are applied to the land-based robots mostly. Furthermore, some of the better formation control strategies stay only in the MATLAB simulation stage.Thus, inspired by the basic behaviors in nature, and supported by the multi-robot formation control techniques existed, this article investigates the formation control of multiple bio-mimetic robotic fish using the robotic fish simulation platform.To realize the formation control of multi-robot fish system, this paper develops a position feedback control strategy, which considers both graph theory and information interaction between the robots. The conception of reference robot fish is introduced, which means a different fish is set to stay still at any time. Other fishes adjust their positions according to the position of reference fish through the information interaction, which can make the robotic fish achieve desired formation in the limit time. The experimental results on the Underwater Robot Polo Game2D Simulation Platform validate the robustness and effectiveness of the proposed algorithms.A novel virtual oval location detecting method is proposed for formation keeping and obstacle avoidance. According to the position of obstacles, the formation deforms locally or transforms entire by the ellipse compression and expansion. The flexible detection method can make the multi-robotic fish system avoid obstacles and restore the original formation rapidly. And the experiments on the simulation platform verify the effectiveness of this approach.