Design And Optimization Of Persistent Monitoring Based On Multiple Agents

The multi-agent persistent monitoring system has been widely used in ocean monitoring and urban patrol due to advantages of its flexible strategy and wide coverage,thus,carrying out the research of multi-agent persistent monitoring has important theoretical value and potential practical value.In view of this,this thesis is based on the first-order and second-order multi-agent systems,and conducts in-depth research on the different monitoring needs in the persistent monitoring problem,such as monitoring different important target points,avoiding collisions among agents and avoiding collisions between agents and obstacles.The main contents of this thesis are as follows:For the one-dimensional persistent monitoring problem of first-order multi-agent,a more general weighted objective function model is established.With the help of Pontryagin Minimum Principle(PMP),the optimal motion law of the agent is obtained.It is found that the agent may stop at the dwelling point and the optimal trajectory can be parameterized.Further,With the help of Infinitesimal Perturbation Analysis(IPA)and a momentum method,the local optimal trajectory is obtained.The simulation results verify the correctness of the theoretical results.For the one-dimensional persistent monitoring problem of second-order multi-agent,under the constraints of no collisions among agents and no boundary crossings,a new objective function is designed based on the external penalty function method.With the help of PMP,the conclusion that the optimal trajectory can be parameterized is obtained.And the local optimal collision-free and crossing-free trajectory is obtained by IPA and a steepest descent method.The simulation results verify the effectiveness of the proposed method of avoiding collisions among agents and boundary crossings.For the two-dimensional persistent monitoring problem,the implementation of persistent monitoring task with second-order multi-agent is proposed for the first time,and the constraints of no collisions among agents and no collisions between agents and obstacles are proposed.With the help of PMP,the optimal motion law of the agent is obtained.On the basis of assigning each agent an elliptical trajectory,the local optimal collision-free elliptical trajectory is obtained by IPA and a steepest descent method.The simulation results verify the effectiveness of the proposed method of avoiding collisions among agents and collisions between agents and obstacles.