移动机器人三维地图创建的仿真研究

Simulation Research of the 3D- Map Creation with A Mobile Robot

利用传感器信号自主创建三维地图是移动机器人智能导航和环境探索研究的基础。利用Blender建模软件建立了不同室外场景的三维模型:在ROS系统中利用C++编程语言设计了飞行器的飞行路线;在Gazebo仿真系统中利用与飞行器连接的激光扫描器对地形进行了扫描;最后利用Octomap地图创建方法生成相应的三维地图。通过对不同参数及传感器连接方式的比较分析,得出结论:飞行器和激光扫描器之间使用刚性连接比铰接更能节省时间和内存;刚性连接的缺点是无法识别被遮盖的地形信息;Octomap方法可以应用到大面积的室外地形。

The use of sensor signals to create 3D-map is the basis for the mobile robot intelligent navigation and environmental exploration studies. In this paper,three-dimensional models of different outdoor scenes are built with the modeling software Blender; then the quadcopter's flight path is designed by using the C + + programming language in the robot operating system; the terrain is scanned by a laser scanner which is mounted under the quadcopter in the Gazebo simulator; finally,a relevant 3D-map is generated by using the Octomap mapping method. By Comparing and analyzing the influence of the different parameters and the sensor connection type,we have reached the following conclusions: compared with the revolute joint between the quadcopter and the laser scanner,the fixed one saves more time and more memory usage; however,the disadvantage of the fixed joint is that the covered topographic information is missing; the Octomap method can be applied to the terrain with a larger area.