Research On Formation Control And Hunting Of Multi-robot Fish Based On Artificial Immune

The increasing depletion of land resources has caused mankind to shift the focus of resource development and competition to the sea,and maritime territorial disputes have gradually become the focus of struggle in international relations.As people pay more and more attention to the fields of underwater detection,target search,underwater demining and anti-submarine,underwater vehicles have developed rapidly.Compared with other underwater robots,robotic fish has the advantages of high mobility and good concealment.Multiple robotic fish line up in a fixed formation to advance,which can greatly improve the efficiency of underwater detection,target search and underwater mine clearance of robotic fish,and the cooperative hunting of targets by multiple robotic fish can also greatly improve underwater anti-submarine.Therefore,this thesis studies the multi-robot system and mainly studies the formation control and cooperative hunting problems in the multi-robot system.The research content of this article is as follows:Firstly,for the three-joint bionic robotic fish,the kinematics modeling is carried out around the control requirements of formation control and cooperative hunting.According to the fish body wave model,the motion state of each joint is analyzed when the bionic robotic fish swims,and the joint motion of the robotic fish is simulated by MATLAB to obtain the control parameters of the robotic fish in different swimming states.Secondly,aiming at the problem of multi-robots deviating from the predetermined trajectory in formation control,an interleukin-regulated multi-robots formation immune network control algorithm is proposed.The algorithm uses the speed of the robotic fish and the center of the expected formation as antibodies and antigens,and draws on Jerne’s idiotype immune network hypothesis,and builds a formation control immune collaboration network through the stimulation and inhibition of antibodies and antigens;inspires the mechanism of interleukin immune regulation,define the interleukin adjustment factor to solve the problem of correcting the robot fish track deviation;the formation control is completed by the adaptive selection speed of the antibody excitation value.Numerical test and simulation platform test results show that compared with other algorithms,the required time is reduced by 72% and 31%,the displacement is reduced by 59% and 37%,and the average deviation angle is close to 0 and 67%,which verifies the effectiveness and superiority of the algorithm.Thirdly,to solve the problem of predicting the movement of hunters in cooperative hunting of multi-robot fish,a cooperative hunting algorithm of multi-robot immune network based on parallel guidance law is proposed.The algorithm uses the movement strategy of the hunter as the antibody,and the location information of the escaper,the target area and the other hunters as the antigen.It draws on Jerne’s idiotype immune network hypothesis,and builds a collaborative hunting immune collaboration network through the stimulation and inhibition of antibodies and antigens;inspired by the guidance law,the parallel guidance law adjustment factor is defined to solve the problem of predicting the movement of the hunter;the hunter adaptively chooses the movement strategy based on the antibody concentration to complete the hunting task.Numerical tests and simulation platform test results show that,compared with other immune algorithms,the immune network cooperative hunting algorithm in the article not only guarantees the success rate of hunting,but also reduces the number of steps required by an average of 23%,the time is reduced by17%,and the corners are reduced by an average of 13 % and 20%,which verifies the effectiveness and superiority of the algorithm.Finally,based on the multi-water robotic fish cooperative control platform,the feasibility verification of multi-robot formation control and cooperative hunting was carried out.In the experiment of multi-water robotic fish cooperative control platform,the time required for the formation control immune network algorithm is reduced by an average of90%,the displacement is reduced by an average of 82%,and the deviation angle is reduced by an average of 42%;the time required for the cooperative hunting immune network algorithm is reduced by an average of 26%,and the turning angle is reduced by 58% on average,further verifying the feasibility and superiority of the algorithm in the article.