curobo.util.warp_interpolation module

get_cuda_linear_interpolation(

raw_traj: JointState,

traj_tsteps: Tensor,

out_traj: JointState,

opt_dt: Tensor,

raw_dt: float = 0.5,

)

Use warp to perform linear interpolation on GPU for a batch of trajectories.

#NOTE: There is a bug in the indexing which makes the last horizon step in the trajectory to be not missed. This will not affect solutions solved by our arm_base class as we make last 3 timesteps the same.

Parameters:

• raw_traj (JointState) – _description_

• traj_tsteps (torch.Tensor) – _description_

• out_traj (JointState) – _description_

• opt_dt (torch.Tensor) – _description_

• raw_dt (float, optional) – _description_. Defaults to 0.5.

Returns:

_description_

Return type:

_type_