ctu-mrs / simulation

The OLD Gazebo Simulation repository.
https://github.com/ctu-mrs/mrs_uav_system
BSD 3-Clause "New" or "Revised" License
48 stars 17 forks source link

Increase GPS random walk to better reflect reality #27

Open klaxalk opened 3 years ago

spurnvoj commented 3 years ago

How much do you want to change it? These are the values which are there now

          <plugin name='gps_plugin' filename='libgazebo_gps_plugin.so'>
            <robotNamespace/>
            <gpsNoise>1</gpsNoise>
            <gpsXYRandomWalk>2.0</gpsXYRandomWalk>
            <gpsZRandomWalk>4.0</gpsZRandomWalk>
            <gpsXYNoiseDensity>0.0002</gpsXYNoiseDensity>
            <gpsZNoiseDensity>0.0004</gpsZNoiseDensity>
            <gpsVXYNoiseDensity>0.2</gpsVXYNoiseDensity>
            <gpsVZNoiseDensity>0.4</gpsVZNoiseDensity>
          </plugin>
spurnvoj commented 3 years ago

Datasheet for Neo-M8N chip used in most of the pixhawk GPS receivers says that the horizontal precision is 2.5 m (GPS+GLONASS). So I can update it to this value. Is it enough?

klaxalk commented 3 years ago

I think 2.5 m is the "optimistic" value. Now the question is, wherher we should not simulate with more pessimistic values, to prevent any surprises in the real world.

spurnvoj commented 1 year ago

Is this is valid? Should I change the GPS horizontal precision to 2.5m?

klaxalk commented 1 year ago

Ye, I still think the normal GPS random walk should be more visible.