0
$\begingroup$

Rosanswers logo

I'm using a Turtlebot 2 with Kinect-sensor, ROS distro is Indigo.

As I understand the sensors IR laser measures a point cloud from which is taken a slice with definable thickness. This slice is then compressed into a 2d plane so that points nearest the robot are taken into account as obstacles.

I tested the thickness of the observation area in gmapping_demo (with default settings) and noticed that it was only a really narrow plane at the height of the sensor. Ignoring obstacles below and above this plane causes the robot e.g. to try to navigate below things it isn't able to go under.

I tried to change two parameters in costmap_common_params.yaml:

scan/min_obstacle_height: from 0.25 to 0.01 and
scan/max_obstacle_height: from 0.35 to 0.50

this didn't have any effect.

Is it possible to adjust the thickness of the "observation layer" - how could I do it?


Originally posted by Erkki on ROS Answers with karma: 3 on 2016-05-10

Post score: 0

$\endgroup$

1 Answer 1

0
$\begingroup$

Rosanswers logo

Hi,

I think you can do this by editing the turtlebot_bringup / 3dsensor.launch file.

In this file, you will find the parameter "scan_height".

Just change the value (10 by default) with the value you want, for example :

<param name="scan_height" value="50"/>

Originally posted by Jbot with karma: 429 on 2016-05-10

This answer was ACCEPTED on the original site

Post score: 1


Original comments

Comment by Erkki on 2016-05-10:
OK, thanks! Unfortunately even if the slice thickness could be increased, the parameter seem to define vertically more or less an arc-type scan area, so it depends also on measuring distance whether an obstacle is marked into the map or not. Those costmap parameters did neither have effect on this.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.