The site is read-only. Please transition to use Robotics Stack Exchange
| ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange |
 | 1 | initial version |
You have the values backward; In the costmap, 255 is occupied, and 0 is free space. Values in between capture the uncertainty about whether a particular cell is occupied or not.
| | 2 | No.2 Revision |
You have the values backward; In the costmap, 255 is occupied, and 0 is free space. Values in between capture the uncertainty about whether a particular cell is occupied or not.
It looks like your costmap rendering is flipped vertically compared to the display in rviz. I suspect this is due to differences in how the axes are aligned in rviz vs images.
The black areas you see correspond to 0 values in the costmap; these areas are free space. This includes all of the space between your laser and the object it's seeing (minus inflation). This also includes the footprint of the robot, because in order for the planner to operate properly, the robot cannot be in a start state that is colliding with anything.
The white areas that you see in the costmap are either occupied or unknown space. Occupied space is marked explicitly by the laser, and unknown space is space where there are no sensor readings, such as behind the robot, and on the far side of obstacles.
| | 3 | No.3 Revision |
You have the values backward; In the costmap, 255 is occupied, and 0 is free space. Values in between capture the uncertainty about whether a particular cell is occupied or not.
It looks like your costmap rendering is flipped vertically compared to the display in rviz. I suspect this is due to differences in how the axes are aligned in rviz vs images.
The black areas you see correspond to 0 values in the costmap; these areas are free space. This includes all of the space between your laser and the object it's seeing (minus inflation). This also includes the footprint of the robot, because in order for the planner to operate properly, the robot cannot be in a start state that is colliding with anything.
The white areas that you see in the costmap are either occupied or unknown space. Occupied space is marked explicitly by the laser, and unknown space is space where there are no sensor readings, such as behind the robot, and on the far side of obstacles.
The obstacle layer in the costmap computes free space by using raytracing to mark the space between the sensor and the obstacle as free. If there are no obstacles in the environment, there's nothing to raytrace to, so all of the space around the robot remains unknown.
ROS Answers is licensed under Creative Commons Attribution 3.0 Content on this site is licensed under a Creative Commons Attribution Share Alike 3.0 license.
