Closed woensug-choi closed 4 years ago
Testing hardcoded tau(external force) as vector6d (1, 0, 0, 0, 0, 0) showed following velocity changes (Capture). The self-drowning characteristic persists.
The original problem lead to this issue.
Insane movements found on UUV simulator's original plugin !!
roslaunch uuv_gazebo_worlds ocean_waves.launch roslaunch eca_a9_description upload_eca_a9.launch
which uploads eca_a9 model of UUV simulator with its plugins. The AUV moves upward gradually when uploaded with neutural buoyancy option turned off. When neuturally_buoyancy option is turned on, it drowns slowly (same problem with modified hybrid glider plugin). HOWEVER, if waited long enough, the model suddenly make very insane movements without any external forces applied. Don't know whats causing it. It must be related to UUV simulator bug '
tag does nothing. #77' which remains an open issue.
<gravity>0</gravity>
right after <physics>
at world file fixes drowning. Need more tests how it effects kinematics.
Following is the initial logs when the model is uploaded to the world in gazebo. Even though initial velocities forces are zeros, suddenly -z directional velocity emerges. It is calculated using 'GetRelativeLinearVel()' which is gazebo's original function. Changing time step schemes in the world file does change rate, but not much. 'ComputeAcc(velRel, _time, 0.3)' at UUV simulator does control oscillation but does not change its tendency to increase (tendency to drown). Maybe it's only causing problem when there's no external forces... (need to construct a case with MATLAB to validate this). Confirmed that no ocean current is set. (all mean values and noise amplitudes are set zero).