Dynamic Coverage Control Of Multi-Agent Systems

In recent years,the problems of distributed cooperative control of multi-agent systems has aroused widespread concern in the academic community.Among them,the coverage control problem with the background of environment monitoring has gradually received the attention of the academic community.Coverage control not only involves agent-agent interactions,but also agent-environment connections.Besides,the final states of agents are needed to be optimized.The theory of coverage control can help us accomplish various complex tasks,such as environmental monitoring,disaster rescue,space exploration,pollution control,etc.Hence,the research of optimal coverage problem has important theoretical significance and application value.Based on the above analysis,for cooperative pollution clearing task,using the tools of computational geometry,distributed optimization and stability theory,this thesis demonstrates and realizes dynamic coverage control of multi-agent systems over the environment under different conditions.The main contents of this thesis are summarized as follows:Focusing on the actuation-based task,dynamic coverage control problem of multi-agent systems is studied.For the existing works of coverage control,researchers mainly consider cooperative monitoring tasks in the static and safe environment of interest.However,the dynamic environment with all kinds of threats and dangers is seldom taken into consideration.When there are harmful pollutants in the environment,agents need to carry actuators to continuously remove the pollutants and reduce the damage to the environment.Compared with cooperative sensing tasks,cooperative clearing tasks must consider the effects of actuators on the distribution of pollutants.To this end,a density function based on the actuator performance function is defined,and an objective function is constructed directly with the average density value of environment.In order to optimize the objective function,a dynamic coverage control algorithm based on state switching is designed.It is proved that the designed control algorithm can ensure that the objective function asymptotically converges to the minimum value.Compared with the existing coverage control methods,the algorithm has better optimization effect and smaller tracking error.Due to the fact that the radius of actuator is limited and the load distribution is unbalanced,dynamic coverage control problem of multi-agent systems based on actuator load balancing is studied.A density function based on the actuator performance function of tunable parameter is defined,and a generalized radius-limited Voronoi partition is constructed.Through computing constrained Voronoi centroid in a distributed manner,the coverage control problem of multi-agent systems is transformed into the target trajectory tracking problem of a single agent.Then the dynamic coverage control algorithm based on state switching is designed and the convergence of the control algorithm is proved.In addition,because the load of each actuator is not the same,in order to achieve the balance of load,the heterogeneous generalized Voronoi partition is constructed and the limiter-based actuator load adjustment method is designed.The simulation results show that the proposed method can effectively avoid multi-agent systems falling into local optimal configuration.Due to the fact that the number,arrival time and orientation of mobile pollution sources are unknown in advance,dynamic coverage control problem of multi-agent systems based on the balance between sensing task and actuating task is studied.In order to find and remove pollution sources timely,a variable of task assignment for each agent is introduced.Based on the variable,the dynamic coverage control problem is transformed into a multi-objective optimization problem with respect to the total travel cost of all agents.Using the Leibniz differential rule of integral,the condition that the objective function takes the minimum value is obtained,and then the stagnation point set of the objective function is determined.The motion controller and task assignment controller of the agent are designed respectively by using the state-switching control method,and the stability of the closed loop system is proved.While ensuring that the objective function converges to the minimum value,the designed control algorithm can significantly improve the success ratio of discovering mobile pollution sources.Considering the limited amount of actuator chemical neutralizers and the distribution of pollutant concentration obeying the diffusion equation,dynamic coverage control problem of multi-agent systems based on reaction-diffusion equation is studied.The spraying process of the actuator is introduced as a feedback control item into the reaction diffusion equation of the pollutant.Through the finite difference method,the distribution function of pollutant concentration is reconstructed and used as the density function of environment.Based on this density function,an objective function is defined for the total travel cost of all agents.In order to minimize the objective function and ensure that the average density value of environment is always below the given value,the motion controller and the diffusion process controller of each agent are designed by using the state-switching control method.As a result,the best removal effect of multi-agent systems on environmental pollutants is achieved.