Gazebo | Ignition | Community
Ask Your Question

Implementation of a Ground Sensing Sensor

asked 2016-08-12 07:50:36 -0500

Illuminatur gravatar image

Hi Guys,

i have a complex geometric file, exported as Collada. The building consists of several layers, such as a multi-layered houses with bumpy levels, I want to create a plugin (for example a sensor plugin), which is able to perceive or calculate the following information, which could be called a Ground Sensing Sensor. The following information shall be derived and calculated:

  • project the sensor position on the nearest layer of the building (corresponds to a ray casting with a plumb vector with z=(0,0,1)
  • Then, the z-coordinate and the normal vector of this projected point shall be printed.

I checked the ContactManager class, which is used in the Contact Sensor Plugin, but i think this is only applicable for contacts, which already have been occurred. I also would like to work with the collision data structures, because i think this is much cheaper than working with a sensor based on depth sensor for example.

Do you have any ideas?

Best regards!

edit retag flag offensive close merge delete

1 Answer

Sort by ยป oldest newest most voted

answered 2016-08-12 10:23:18 -0500

nkoenig gravatar image

You could create a laser sensor with a single beam. Then orient the sensor to face down. The laser will return the distance from the sensor to the nearest contact.

edit flag offensive delete link more


Hi, thanks for your quick response. I think, using a sensor this way, would cause wrong results, since i need an exact projection on the ground plane of the current level. A sensor's pitching movement would then also result in a wrong distance value. Is it possible to access the geometric collision model of a dae inside gazebo's data structure?

Illuminatur gravatar imageIlluminatur ( 2016-08-12 11:15:13 -0500 )edit

How does Gazebo handle the collada-based collision models? In which data structures are they stored and can these structures be accessed within a sensor or model plugin? I think such an approach would be the most elegant. Best regards and thanks a lot!

Illuminatur gravatar imageIlluminatur ( 2016-08-15 02:49:00 -0500 )edit

Question Tools



Asked: 2016-08-12 07:50:36 -0500

Seen: 439 times

Last updated: Aug 12 '16