Research Of Mobile Robot Navigation Based On ROS In Unknown Environment

Since rescue robots can replace people to get to dangerous fields such as fires and finish serving as well as cleaning in restaurants and families,it attracts broad attention from domestic and overseas research institutes.The mobile robot in the unknown environment has only a little prior knowledge,and the navigation control method involves many key issues such as environmental cognition,behavior decision,motion control and so on.In this paper,the navigation system is constructed by ROS,and it has positive scientific value and practical significance.Before autonomous navigation,the mobile robot needs the environment map of the unknown area,therefore,it needs to construct the map of the environment.Aiming at the problem of large error in the complex environment,this paper totally depends on the distance information of the laser without the need of additional sensor data.In the actual environment the grid map of indoor corridor construction achieves good experimental results.In order to solve the problem of "where am I" in the navigation process,we must realize real-time global positioning based on the known environment model.In this paper,the weight is sampled and calculated by the particle filter,and the distance information is collected by laser radar,The prediction of the distance information and the recursive cycle of resampling is updated.In the actual experiment,the conversion relationship among the world coordinate system?inertia coordinate system and the robot chassis frame is given,then the global pose of the mobile robot is obtained.Mobile robot path planning problem is the key to realize autonomous navigation,this paper according to the needs of path planning in a dynamic environment proposes the ways of combining global planning and local planning algorithm,and simulation experiments are analyzed.When it comes to global path planning algorithm,this paper uses improved RRT*algorithm based on RRT.Not only could probabilistic completeness of algorithm guarantee,but also the planning effect is optimal probability through adding the step of triming the tree structure.At the same time,the local path planning DWA algorithm is also discussed,which can get the optimal path from the sample trajectory in the dynamic environment.The combination of global and local planning can make use of the known map information as well as meet the real-time requirements so as to achieve better planning effect.In the aspect of hardware,the mobile robot platform is made up of iRobot Create differential driving chassis.UTM 30LX laser radar and PC etc.Besides,the building of navigation framework and the related algorithm implementation are completed with the help of navigation stack in ROS system.At the same time,the results of the fixed point navigation show that the mobile robot platform has good robustness,and can effectively avoid the dynamic obstacles in the procession to reach the target.In addition,the final attempts of sending and receiving commands between the remote terminal and the mobile robot through the wireless sensor link also achieve good experimental results.