You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I want to ask a question how to use integrate parallel trajectory generation to Isaac lab.
Motivation:
Curobo has feature of planning in multiple environments. If this can be seamlessly integrate into isaac sim parallel environment, the benefit for generating quality data is huge. However, it seems like how to use it is not directly clear.
Context:
Curobo provides planning function as such : trajectories = plan_batch(num_envs, envs), and the output trajectories will be list of dimension [num_envs, N_i] where N_i may different for each environment.
The only constraint is that it num_envs always needs to consistent.
Trouble:
When Trajectory planning is used in a batch environment, one possible troublesome scenario is when a robot that has done execution has to wait for another robot that hasn’t done its execution. This is undesired because each environment reset time depends on other environments' termination conditions.
Example:
Say if we have 2 environments.
Robot1 planned trajectory length is 45
Robot2 planned trajectory length is 100
if we step once for both environment, clearly Robot1 finishes trajectory before Robot2, while Robot2 keeps stepping its trajectory, the Robot1 must wait until Robot2 done before next time plan_batch(2, envs) is called. (Remember plan_batch always receives consistent num_envs, so it must plan both environment at the same time)
Plausible Solution:
If isaac sim or isaac lab can specify which environment to step and which not to step, then this problem can be solved quite easily, however due to my limited knowledge it is not immediately clear how to do that. Or do you have any other suggestion?
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
I want to ask a question how to use integrate parallel trajectory generation to Isaac lab.
Motivation:
Curobo has feature of planning in multiple environments. If this can be seamlessly integrate into isaac sim parallel environment, the benefit for generating quality data is huge. However, it seems like how to use it is not directly clear.
Context:
Curobo provides planning function as such : trajectories = plan_batch(num_envs, envs), and the output trajectories will be list of dimension [num_envs, N_i] where N_i may different for each environment.
The only constraint is that it num_envs always needs to consistent.
Trouble:
When Trajectory planning is used in a batch environment, one possible troublesome scenario is when a robot that has done execution has to wait for another robot that hasn’t done its execution. This is undesired because each environment reset time depends on other environments' termination conditions.
Example:
Say if we have 2 environments.
Robot1 planned trajectory length is 45
Robot2 planned trajectory length is 100
if we step once for both environment, clearly Robot1 finishes trajectory before Robot2, while Robot2 keeps stepping its trajectory, the Robot1 must wait until Robot2 done before next time
plan_batch(2, envs)
is called. (Remember plan_batch always receives consistent num_envs, so it must plan both environment at the same time)Plausible Solution:
If isaac sim or isaac lab can specify which environment to step and which not to step, then this problem can be solved quite easily, however due to my limited knowledge it is not immediately clear how to do that. Or do you have any other suggestion?
Beta Was this translation helpful? Give feedback.
All reactions