ROS 自定义地图加载和导航（turtlebot测试）

1、构建地图

• 启动turtlebot底盘

roslaunch turtlebot_bringup minimal.launch 

• 启动传感器（激光雷达和惯性导航）

roslaunch lms1xx LMS1xx.launch #启动激光雷达

• 启动hector slam

roslaunch lms1xx hector_mapping_demo.launch

• 启动远程遥控

roslaunch turtlebot_teleop keyboard_teleop.launch

• 启动遥控xbox

  roslaunch turtlebot_teleop xbox360_teleop.launch
  ​
  ​
  #(1)
  #(2) joy_node节点修改
  #
  #   
  #   
  #  
  

• 保存地图

rosrun map_server map_saver -f ~/my_map

2、自定义地图amcl导航

• 启动turtlebot底盘

roslaunch turtlebot_bringup minimal.launch

• 启动激光雷达

roslaunch turtlebot_navigation lms1xx.launch 

• 启动amcl导航

roslaunch turtlebot_navigation lab_amcl_nomap.launch #不带加载地图，单独加载

 

• 加载地图

rosrun map_server map_server my_map.yaml

• rviz查看

  roslaunch turtlebot_navigation view_robot.launch 

   

3、问题

• （1）地图显示不正确

打开yaml文件

image: my_map.png
resolution: 0.050000
origin: [-51.224998, -51.224998, 0.000000]
negate: 0
occupied_thresh: 0.65
free_thresh: 0.196

注意：

图像像素值[0-255].

• If negate is false, p = (255 - x) / 255.0. This means that black (0) now has the highest value (1.0) and white (255) has the lowest (0.0).

• If negate is true, p = x / 255.0. This is the non-standard interpretation of images, which is why it is called negate, even though the math indicates that x is not negated. Nomenclature is hard.

一般negate为0。

p = (255 - x) / 255.0

• If p > occupied_thresh（0.65，换算为灰度值，约为90，也就是说像素小于90为占用（障碍物））, output the value 100 to indicate the cell is occupied.

• If p < free_thresh（0.196，换算为灰度值，约为205，也就是说像素大于205为空（自由空间站））, output the value 0 to indicate the cell is free.

• Otherwise, output -1 a.k.a. 255 (as an unsigned char), to indicate that the cell is unknown.

我通过hector保存的图像自由空间的值为200，显示为障碍物，所以出现上面图像错误。