Research On Custom Obstacle Layers Of Mobile Robot Costmap

The navigation function of the mobile robot is the prerequisite for realizing all its value,and the map is the basis for determining whether the robot can navigate smoothly.At present,most indoor handling robots and inspection robots use layered costmaps for navigation.The general layered costmap is composed of a static map layer,an obstacle layer,and an expansion layer.They can meet the basic requirements of the mobile robot navigation process.However,in the actual application process,the use of mobile robots,they often face some new problems.These problems may not be solved by using the existing functional layers.In this thesis,a custom barrier layer is designed.This layer can add virtual obstacles to the costmap freely through the artificial information,and is not affected by the external environment.The virtual obstacles can limit the robot’s activity area and change the robot’s moving path,so as to affect the robot’s navigation.This thesis takes the logistics scene as the research background,and takes the mobile robot cost map as the main research object to do the following research.Firstly,the mathematical model is established for the sensors used by the mobile robot,and the navigation model of the mobile robot under the ROS platform is established.Based on the analysis and simulation of the path planning algorithm represented by A* algorithm,the problems existing in the global path planning of some common path planning algorithms of mobile robots are introduced,that is,the planned path is too close to the obstacle expansion area at the inflection point.This problem can easily cause the robot to cross the expansion area and hit obstacles because of the cumulative error when turning.This thesis proposes a method to add virtual obstacles at the corner to solve the problem.Secondly,the common costmaps are studied,some common costmaps function layers are studied,and the new layer routine Simple layer provided by the ROS platform is studied.Based on the Simple layer,a custom obstacle layer that can add virtual obstacles freely in the cost map is designed.Compared with other layers,this layer can be widely used in various scenes,and can artificially give virtual obstacle information.Thirdly,taking the logistics scene as the research background,the traditional Harris corner detection algorithm is improved by using the "minimum circumscribed rectangle" algorithm.The improved Harris algorithm has the ability to accurately detect the corners of obstacles in the static map.Using this algorithm can provide the virtual obstacle coordinate information to be added to the custom obstacle layer.Finally,a real mobile robot platform Hi Bot and a set of experimental scenes were used to verify the research content in this thesis.Through experiments,we found that the virtual obstacles added with the custom obstacle layer can effectively affect the global path planning and real trajectory of the mobile robot.The improved Harris algorithm can accurately detect the inflection point of the obstacle in the static map.Mobile robots use costmaps after adding virtual obstacles to navigate.The trajectory is smoother when turning and keeping a larger distance from real obstacles,greatly reducing the possibility of collisions.The experimental results prove the authenticity,reliability and practicability of the research content.