Research On Persistent Monitoring Algorithm Of Multi-Robot System

In the existing literature,the research,about the multi-robot persistent monitoring task,mainly focuses on the issue that given a closed path.The multi-robot is performed along the given path to achieve a desired target,when performs a task.This paper optimizes the multi-robot persistent monitoring algorithm,so that the robot can independently plan the path to complete the persistent monitoring task according to the relevant information of the target point.The basic idea is that the performed persistent monitoring robot can use the information,which expresses the degree of access of the node.The robot uses the information,such as idleness time,node uncertainty and other indicators,to make the next decision to access the node.At the same time,the information is also used as an evaluation indicator for global tasks.The main research contents and achievements include the following aspects:(1)From the covering problem in the wireless sensor network to the multi-robot persistent monitoring problem,the author studies the robot perception model for the multi-robot persistent monitoring algorithm,by using the Voronoi partitioning and the virtual force algorithm at the same time.In addition to introducing a more realistic robot sensing model,the virtual force is used to make the robot move autonomously from the initial position to the final optimized position,which ultimately maximizes the coverage of the robot's persistent monitoring of the mission area.Finally,by comparing the proposed algorithm with the simulation results of the 0-1 perceptual model,the correctness and effectiveness of the improved algorithm are proved.(2)Studying the problem of effective evaluation criteria for persistent monitoring of multiple robots and the path of robot autonomous selection.In view of the existing literature,the evaluation criteria,used in the persistent monitoring tasks of multi-robot systems,are mostly linear.And during the execution of the mission,the robots are moving along a given closed path.Here will adopt the uncertainty target point,which changed nonlinearity with time,as the evaluation index of the persistent monitoring task,and using the Dynamic Windows Approach to plan the path independently.When the robot uses the indicator to select the next target point,the uncertainty value of the global target point is minimized.The effectiveness of the algorithm is verified by a ROS-based simulation experiment platform.(3)In the current research on the persistent monitoring of multi-robots,most of the target points are regarded as equally important,but it is not actually.The target points in the mission area are divided into important target nodes and secondary target nodes to better allocate robot resources.Before the task starts,the importance degree of the target node is determined by different prior probabilities,and the idle time is used as the evaluation index of the system.The number of times of the important target nodes are accessed higher than the number of the secondary target nodes.The simulation experiment verifies the feasibility and effectiveness of the optimized algorithm.Obviously,the proposed algorithm is more in line with actual needs.