Research On Cooperative Control Of Multiple Intelligent Biomimetic Robotic Fish

Rencently, the cooperative control of multiple underwater bio-robots have been widelystudied as the human requirement increases in the fields of exploration of marine resourceswater environment detection and military, etc. Generally speaking, it is difficult to control thebiomimetic robotic fish or its group precisely since the underwater environment is complexand the precise dynamics of robotic fish is unknown. This thesis considers the formationcontrol, object transportation and accurate posture control problems. The main contents are asfollows.1. Two path tracking control algorithms are proposed for precise posture control ofbiomimetic robotic fish with nonholonomic constraints. In the first one, the fuzzy controller isdesigned to control the linear and angular velocity of robotic fish and thus the accuratetracking is achieved. In the second one, the linear velocity is calculated by fuzzy controller,and then the desired path and the angular velocity are obtained by Bezier curve method.2. Three hierarchical control schemes are proposed for multi-robotic fish with fixedtopology to form and maintain the formation.1) Base on the consensus theory, a distributedalgorithm is first proposed to estimate the reference point of formation, and then anotherconsensus protocol is given to achieve the accurate tracking.2) Just as the above, thereference point of formation is estimated by a distributed consensus algorithm. The differentlies that the controller based on individual is nonlinear. The accurate tracking is also achieved.3) In this framework, the algorithm used to estimate the reference point is just the same as thesecond method. The different lies that individual-based controller is fuzzy.3. Two distributed strategies are presented for multi-robotic fish to transport the objectcooperatively.1) In this scheme，the box is taken as a leader and the robotic fish followers.The synchronization of all robotic fish is guaranteed by consensus algorithm and then theposture and pushing force of each individual is decided by the posture of box, whichcoordinates all robotic fish to achieve the task.2) A distributed control algorithm is proposedto convert cooperative transportation task into a formation control problem, where the box isthe leader. For each robotic fish, the Bezier curve and the fuzzy controller are used to controlthe angular velocity and linear velocity, respectively. The presented method not onlyguarantees the synchronization of multi-robotic fish, but also achieves the distributedcooperative box-pushing task.