Research On Swarm-based Formation Control And System Implementation Based On Bionic Cluster Model

With the development of modern marine ranching,the assessment of fishery resources has become increasingly important.Currently,fishery resource assessment mainly relies on fishing sampling or diver underwater observation methods,which have many limitations such as poor safety and rough assessment results.With the development of underwater robot technology,individual autonomous underwater robots are widely used in fishery resource assessment.However,the assessment coverage is extensive,therefore,multiple underwater robot systems with advantages of spatiotemporal distribution and high efficiency have become a potential effective means of fishery resource assessment.Formation control is the most basic task in multi-robot systems.Compared with traditional control strategies that require the design of complex dynamic models,extensive parameter optimization,and high computational complexity,bio-inspired formation control strategies have advantages in the limited computing and motion capabilities of underwater robots and complex underwater environments.This paper focuses on the research of bio-inspired formation control methods and the constraints of underwater robots in real environments.The research content includes:(1)A formation control algorithm that combines the social force model with the Boids model.The proposed algorithm addresses the issue of traditional formation control strategies requiring the design of complex dynamic models and extensive parameter optimization.Based on the traditional social force model,the algorithm designs a social force model for individual agents in a formation of underwater robots and incorporates ideas from the Boids model.Specifically,the algorithm combines the concept of three regions into the design of the social force model and divides the forces for different regions to achieve the final formation control.Experimental results demonstrate that this algorithm does not require complex models and can achieve efficient area search tasks and formation maintenance through local interactions.It has the advantages of a simple model,the ability to easily modify the range of forces to change the density of the formation,and is more advantageous for different scene applications.(2)A Minimalistic Formation Control Algorithm Based on Simple Strategy.Inspired by the behavior of fish schools in biomimetics,a simple method for creating collective behavior is proposed to address the problems of the existing methods,such as the need for artificially presetting correlation coefficients and the complexity and difficulty of understanding the calculations.First,the influence of neighboring individuals in the flock is considered,and individual-neighboring coordination is achieved through a certain angle of rotation to the left and right.Second,the influence of a target point on an individual is introduced,such that the individual is attracted to the target point and moves in a straight line towards it,while also interacting with neighboring individuals.Finally,an optimization strategy for individual-target points is developed,which results in the final formation control algorithm.Experimental results show that the extremely simple formation control model proposed in this paper can achieve coordinated motion of the flock through limited angles of rotation to the left and right,and can efficiently complete area search tasks while maintaining flock formation.The algorithm has low requirements for robots and is more advantageous in the subsequent applications of real robots.(3)Design and implementation of a formation control system for area coverage tasks.Due to the development and complexity issues of underwater robots,most large-scale underwater robot formation control methods are currently validated through numerical simulation experiments.Although many realistic environmental constraints are considered,there is still a significant gap between the simulation and real-world environments.Therefore,this paper combines the marine environment and a real underwater robot model to build a formation control system with a realistic physical model.The system integrates sensors,drivers,and other real robot components of simulated robots,allowing for the verification,optimization,and adjustment of algorithms in a relatively realistic environment,while providing real-time monitoring of the status of each robot in the formation.