Research And Implementation Of Virtualand Real Collaborative Navigation System For Robot Clusters Based On ROS

In recent years,with the development of robot technology and artificial intelligence,intelligent robots are more and more used in life.At the same time,it is generally believed that the cooperative work efficiency of robot clusters is higher than that of a single robot.For example,in the aspect of system fault tolerance,the fault tolerance rate of multi-robot cooperative work is very high,and the whole system will not collapse because of the error or damage of a single robot,and more complex tasks can be completed by using robot clusters to cooperate.ROS is one of the most popular open source platforms in the field of robotics,which can effectively package the underlying robot hardware and provide a large number of open source tools and advanced algorithms.In this paper,based on ROS system,the virtual-real collaborative navigation system of robot cluster is deeply studied,which mainly includes the following contents:(1)Firstly,the current mainstream path planning algorithms are studied in detail,and simulation experiments are carried out.It is found that A* algorithm can quickly give the approximate optimal solution to reach the target point based on static map,but for dynamic environment,it is necessary to avoid obstacles through repeated re-planning.In this paper,A* algorithm and DWA algorithm are integrated.The global planner gives the global route to the target point through A* algorithm,and the local planner divides the global route into several local target points.Then DWA algorithm is used to plan according to the global route and local information,and output speed instructions.(2)The A* algorithm is applied to multi-robot collaborative planning problem,and the simulation experiment shows that the shortest path to the target point is planned by a single robot,which often leads to path congestion and frequent conflicts between robots,wasting a lot of time and reducing the overall performance of the system.Inspired by this,this paper proposes an improved A* algorithm based on time window,which assigns a direction time window and an occupation time window to each node on the raster map,records the current running path in the form of time window,updates the time window information of the nodes on the path when planning the path,and eliminates the nodes with direction conflict or occupation time conflict.Modify the heuristic function of A*algorithm,and increase the cost of the impact of nodes on the overall system performance when searching for nodes.(3)Based on ROS and Gazebo,this paper builds a virtual-real combination platform,carries out experiments of virtual-real cooperation of robot clusters,and verifies the effectiveness of multi-robot path planning algorithm.The robot simulation model in which the actual robot is mapped on the virtual environment and the map simulation model matching with the real environment are built.Collaborative control and communication between virtual and real robots are realized.Through the fusion of virtual and real data,the robot cluster can perceive the real environment and virtual simulation environment at the same time.It achieves the effect of mutual perception and cooperative work between virtual and real robots.