storm_kit.mpc.model.urdf_kinematic_model module¶
-
class
URDFKinematicModel
(urdf_path, dt, batch_size=1000, horizon=5, tensor_args={'device': 'cpu', 'dtype': torch.float32}, ee_link_name='ee_link', link_names=[], dt_traj_params=None, vel_scale=0.5, control_space='acc')[source]¶ Bases:
storm_kit.mpc.model.model_base.DynamicsModelBase
-
_abc_impl
= <_abc_data object>¶
-
get_next_state
(curr_state: torch.Tensor, act: torch.Tensor, dt)[source]¶ Does a single step from the current state Args: curr_state: current state act: action dt: time to integrate Returns: next_state
-
rollout_open_loop
(start_state: torch.Tensor, act_seq: torch.Tensor, dt=None) → Tuple[torch.Tensor, torch.Tensor, torch.Tensor][source]¶
-
tensor_step
(state: torch.Tensor, act: torch.Tensor, state_seq: torch.Tensor, dt=None) → torch.Tensor[source]¶ Args: state: [1,N] act: [H,N] todo: Integration with variable dt along trajectory
-