newton.actuators#
GPU-accelerated actuator models for physics simulations.
This module provides a modular library of actuator components — drives,
clamping, and delay — that compute joint effort from simulation state and
control targets. Components are composed into an Actuator instance
and registered with add_actuator() during model
construction.
Experimental
The actuator API may change without prior notice. Feedback is welcome — please file issues or discussion threads.
Classes
Composed actuator: delay → drive → clamping. |
|
Result of parsing a USD actuator prim. |
|
Base class for actuator output effort clamping. |
|
DC motor four-quadrant effort-speed saturation. |
|
Symmetric clamp on actuator output effort. |
|
Position-dependent effort clamping via lookup table. |
|
Classification of actuator component schemas. |
|
Per-DOF command input delay for actuators. |
|
Base class for actuator drives. |
|
LSTM-based neural network actuator drive. |
|
MLP-based neural network actuator drive. |
|
Stateless proportional-derivative actuator drive. |
|
Stateful proportional-integral-derivative actuator drive. |
|
Effective inverse-mass response for each articulation. |
|
Canonical USD tokens from |
Functions
|
Parse a USD Actuator prim into a composed actuator specification. |
|
Register a USD API schema for actuator parsing. |
Deprecated
Name |
Guidance |
|---|---|
|
Deprecated in 1.6; use ClampingBase instead. |
|
Deprecated in 1.6; use DriveBase instead. |
|
Deprecated in 1.6; use DriveNeuralLSTM instead. |
|
Deprecated in 1.6; use DriveNeuralMLP instead. |
|
Deprecated in 1.6; use DrivePD instead. |
|
Deprecated in 1.6; use DrivePID instead. |