MJX API

Public API for MJX.

step(m: Model, d: Data) → Data

Advance simulation.

class Model

Static model of the scene that remains unchanged with each physics step.

nq

number of generalized coordinates

Type:

int

nv

number of degrees of freedom

Type:

int

nu

number of actuators/controls

Type:

int

na

number of activation states

Type:

int

nbody

number of bodies

Type:

int

njnt

number of joints

Type:

int

ngeom

number of geoms

Type:

int

nsite

number of sites

Type:

int

ncam

number of cameras

Type:

int

nlight

number of lights

Type:

int

nmesh

number of meshes

Type:

int

nmeshvert

number of vertices for all meshes

Type:

int

nmeshnormal

number of normals in all meshes

Type:

int

nmeshtexcoord

number of texcoords in all meshes

Type:

int

nmeshface

number of faces for all meshes

Type:

int

nmeshgraph

number of ints in mesh auxiliary data

Type:

int

nmeshpoly

number of polygons in all meshes

Type:

int

nmeshpolyvert

number of vertices in all polygons

Type:

int

nmeshpolymap

number of polygons in vertex map

Type:

int

nhfield

number of heightfields

Type:

int

nhfielddata

size of elevation data

Type:

int

ntex

number of textures

Type:

int

ntexdata

size of texture data

Type:

int

nmat

number of materials

Type:

int

npair

number of predefined geom pairs

Type:

int

nexclude

number of excluded geom pairs

Type:

int

neq

number of equality constraints

Type:

int

ntendon

number of tendons

Type:

int

nwrap

number of wrap objects in all tendon paths

Type:

int

nsensor

number of sensors

Type:

int

nnumeric

number of numeric custom fields

Type:

int

ntuple

number of tuple custom fields

Type:

int

nkey

number of keyframes

Type:

int

nmocap

number of mocap bodies

Type:

int

nM

number of non-zeros in sparse inertia matrix

Type:

int

nB

number of non-zeros in B matrix

Type:

int

nC

number of non-zeros in C matrix

Type:

int

nD

number of non-zeros in D matrix

Type:

int

nJmom

number of non-zeros in Jacobian momentum matrix

Type:

int

nJten

number of non-zeros in sparse tendon Jacobian

Type:

int

ngravcomp

number of bodies with nonzero gravcomp

Type:

int

nuserdata

number of elements in userdata

Type:

int

nsensordata

number of elements in sensor data vector

Type:

int

npluginstate

number of plugin state values

Type:

int

nhistory

number of history buffer elements

Type:

int

opt

physics options

Type:

mujoco.mjx._src.types.Option

stat

model statistics

Type:

mujoco.mjx._src.types.Statistic | mujoco.mjx._src.types.StatisticWarp

qpos0

qpos values at default pose

Type:

jax.Array

qpos_spring

reference pose for springs

Type:

jax.Array

bind(obj: MjsBody | MjsFrame | MjsGeom | MjsJoint | MjsLight | MjsMaterial | MjsSite | MjsMesh | MjsSkin | MjsTexture | MjsText | MjsTuple | MjsCamera | MjsFlex | MjsHField | MjsKey | MjsNumeric | MjsPair | MjsExclude | MjsEquality | MjsTendon | MjsSensor | MjsActuator | MjsPlugin | Iterable[MjsBody | MjsFrame | MjsGeom | MjsJoint | MjsLight | MjsMaterial | MjsSite | MjsMesh | MjsSkin | MjsTexture | MjsText | MjsTuple | MjsCamera | MjsFlex | MjsHField | MjsKey | MjsNumeric | MjsPair | MjsExclude | MjsEquality | MjsTendon | MjsSensor | MjsActuator | MjsPlugin]) → BindModel

Bind a Mujoco spec to an MJX Model.

class Data

Dynamic state that updates each step.

time

simulation time

Type:

jax.Array

qpos

position

Type:

jax.Array

qvel

velocity

Type:

jax.Array

act

actuator activation

Type:

jax.Array

history

actuator history buffer

Type:

jax.Array

qacc_warmstart

warm start for solver

Type:

jax.Array

plugin_state

plugin state values

Type:

jax.Array

ctrl

control input

Type:

jax.Array

qfrc_applied

applied generalized force

Type:

jax.Array

xfrc_applied

applied Cartesian force/torque

Type:

jax.Array

eq_active

enable/disable equality constraints

Type:

jax.Array

mocap_pos

positions of mocap bodies

Type:

jax.Array

mocap_quat

orientations of mocap bodies

Type:

jax.Array

qacc

acceleration

Type:

jax.Array

act_dot

time-derivative of actuator activation

Type:

jax.Array

userdata

user data

Type:

jax.Array

sensordata

sensor data output

Type:

jax.Array

xpos

Cartesian position of body frame

Type:

jax.Array

xquat

Cartesian orientation of body frame

Type:

jax.Array

xmat

rotation matrix of body frame

Type:

jax.Array

xipos

Cartesian position of body com

Type:

jax.Array

ximat

rotation matrix of body inertia

Type:

jax.Array

xanchor

Cartesian position of joint anchor

Type:

jax.Array

xaxis

Cartesian joint axis

Type:

jax.Array

ten_length

tendon lengths

Type:

jax.Array

geom_xpos

Cartesian position of geoms

Type:

jax.Array

geom_xmat

rotation matrix of geoms

Type:

jax.Array

site_xpos

Cartesian position of sites

Type:

jax.Array

site_xmat

rotation matrix of sites

Type:

jax.Array

cam_xpos

camera positions

Type:

jax.Array

cam_xmat

camera rotation matrices

Type:

jax.Array

subtree_com

com of each subtree

Type:

jax.Array

cvel

center of mass based velocity

Type:

jax.Array

cdof

center of mass based jacobian

Type:

jax.Array

cdof_dot

time-derivative of cdof

Type:

jax.Array

qfrc_bias

C(qpos,qvel)

Type:

jax.Array

qfrc_gravcomp

gravity compensation term

Type:

jax.Array

qfrc_fluid

fluid drag and buoyancy forces

Type:

jax.Array

qfrc_passive

passive force

Type:

jax.Array

qfrc_actuator

actuator force

Type:

jax.Array

actuator_force

actuator force in actuation space

Type:

jax.Array

actuator_length

actuator lengths

Type:

jax.Array

qfrc_smooth

smooth dynamics force

Type:

jax.Array

qacc_smooth

acceleration without constraints

Type:

jax.Array

qfrc_constraint

constraint force

Type:

jax.Array

qfrc_inverse

net external force for inverse dynamics

Type:

jax.Array

bind(model: Model, obj: MjsBody | MjsFrame | MjsGeom | MjsJoint | MjsLight | MjsMaterial | MjsSite | MjsMesh | MjsSkin | MjsTexture | MjsText | MjsTuple | MjsCamera | MjsFlex | MjsHField | MjsKey | MjsNumeric | MjsPair | MjsExclude | MjsEquality | MjsTendon | MjsSensor | MjsActuator | MjsPlugin | Iterable[MjsBody | MjsFrame | MjsGeom | MjsJoint | MjsLight | MjsMaterial | MjsSite | MjsMesh | MjsSkin | MjsTexture | MjsText | MjsTuple | MjsCamera | MjsFlex | MjsHField | MjsKey | MjsNumeric | MjsPair | MjsExclude | MjsEquality | MjsTendon | MjsSensor | MjsActuator | MjsPlugin]) → BindData

Bind a Mujoco spec to an MJX Data.

collision(m: Model, d: Data) → Data

Collides geometries.

refit_bvh(m: Model, d: Data, ctx: Any)

Refit the scene BVH for the current pose.

make_constraint(m: Model, d: Data) → Data

Creates constraint jacobians and other supporting data.

deriv_smooth_vel(m: Model, d: Data) → Array | None

Analytical derivative of smooth forces w.r.t. velocities.

euler(m: Model, d: Data) → Data

Euler integrator, semi-implicit in velocity.

forward(m: Model, d: Data) → Data

Forward dynamics.

fwd_acceleration(m: Model, d: Data) → Data

Add up all non-constraint forces, compute qacc_smooth.

fwd_actuation(m: Model, d: Data) → Data

Actuation-dependent computations.

fwd_position(m: Model, d: Data) → Data

Position-dependent computations.

fwd_velocity(m: Model, d: Data) → Data

Velocity-dependent computations.

implicit(m: Model, d: Data) → Data

Integrates fully implicit in velocity.

rungekutta4(m: Model, d: Data) → Data

Runge-Kutta explicit order 4 integrator.

inverse(m: Model, d: Data) → Data

Inverse dynamics.

create_render_context(mjm: MjModel, nworld: int, devices: Sequence[str] | None = None, **kwargs)

Creates a render context.

Parameters:

• mjm – the MuJoCo model

• nworld – number of worlds to render. We must hardcode the nworld because Warp creates arrays of size nworld that are not exposed to JAX. Thus we cannot use JAX transforms like vmap with the render context.

• devices – optional list of device names (e.g. [‘cuda:0’, ‘cuda:1’]). If provided, rendering workloads are sharded across these devices. By default, devices is None and the default device from wp.get_device(None) is used.

• **kwargs – forwarded to the render context constructor.

Returns:

Render context object that is JAX compatible.

get_data(m: MjModel, d: Data, keepalive_refs: Dict[int, Any] | None = None) → MjData | List[MjData]

Gets mjx.Data from a device, resulting in mujoco.MjData or List[MjData].

get_data_into(result: MjData | List[MjData], m: MjModel, d: Data, keepalive_refs: Dict[int, Any] | None = None)

Gets mjx.Data from a device into an existing mujoco.MjData or list.

get_state(m: Model, d: Data, spec: int | mjtState) → Array

Gets state from mjx.Data. This is equivalent to mujoco.mj_getState.

Parameters:

• m – model describing the simulation

• d – data for the simulation

• spec – int bitmask or mjtState enum specifying which state components to include

Returns:

a flat array of state values

make_data(m: Model | MjModel, device: Device | None = None, impl: str | Impl | None = None, _full_compat: bool = False, nconmax: int | None = None, naconmax: int | None = None, naccdmax: int | None = None, njmax: int | None = None, keepalive_refs: Dict[int, Any] | None = None) → Data

Allocate and initialize Data.

Parameters:

• m – the model to use

• device – which device to use - if unspecified picks the default device

• impl – implementation to use (‘jax’, ‘warp’)

• nconmax – maximum number of contacts to allocate for warp across all worlds Since the number of worlds is not pre-defined in JAX, we use the nconmax argument to set the upper bound for the number of contacts across all worlds. In MuJoCo Warp, the analgous field is called naconmax.

• naconmax – maximum number of contacts to allocate for warp across all worlds Since the number of worlds is not pre-defined in JAX, we use the naconmax argument to set the upper bound for the number of contacts across all worlds, rather than the nconmax argument from MuJoCo Warp.

• naccdmax – maximum number of contacts for GJK collision detection across all worlds. Since the number of worlds is not pre-defined in JAX, we use the naccdmax argument to set the upper bound for the number of contacts across all worlds, rather than the nccdmax argument from MuJoCo Warp.

• njmax – maximum number of constraints to allocate for warp across all worlds

• keepalive_refs – optional dict to store references to underlying MuJoCo objects, preventing them from being garbage collected. Required for CPP impl when passing a types.Model.

Returns:

an initialized mjx.Data placed on device

Raises:

• ValueError – if the model’s impl does not match the make_data impl

• NotImplementedError – if the impl is not implemented yet

• DeprecationWarning – if nconmax is used

put_data(m: MjModel, d: MjData, device: Device | None = None, impl: str | Impl | None = None, nconmax: int | None = None, naconmax: int | None = None, njmax: int | None = None, dummy_arg_for_batching: Array | None = None, keepalive_refs: Dict[int, Any] | None = None) → Data

Puts mujoco.MjData onto a device, resulting in mjx.Data.

Parameters:

• m – the model to use

• d – the data to put on device

• device – which device to use - if unspecified picks the default device

• impl – implementation to use (‘jax’, ‘warp’)

• nconmax – maximum number of contacts to allocate for warp

• naconmax – maximum number of contacts to allocate for warp across all worlds Since the number of worlds is not pre-defined in JAX, we use the naconmax argument to set the upper bound for the number of contacts across all worlds, rather than the nconmax argument from MuJoCo Warp.

• njmax – maximum number of constraints to allocate for warp

• dummy_arg_for_batching – dummy argument to use for batching in cpp implementation

• keepalive_refs – optional dict to store references to underlying MuJoCo objects, preventing them from being garbage collected.

Returns:

an mjx.Data placed on device DeprecationWarning: if nconmax is used

put_model(m: MjModel, device: Device | None = None, impl: str | Impl | None = None, graph_mode: GraphMode | None = None, keepalive_refs: Dict[int, Any] | None = None) → Model

Puts mujoco.MjModel onto a device, resulting in mjx.Model.

Parameters:

• m – the model to put onto device

• device – which device to use - if unspecified picks the default device

• impl – implementation to use

• graph_mode – CUDA graph capture mode (for Warp only). Use GraphMode enum from warp._src.jax_experimental.ffi. GraphMode.WARP is the default mode.

• keepalive_refs – optional dict to store references to underlying MuJoCo objects, preventing them from being garbage collected. Required for CPP impl to keep the model alive.

Returns:

an mjx.Model placed on device

Raises:

• ValueError – if impl is not supported

• RuntimeError – if impl is WARP and warp-lang is not installed

set_state(m: Model, d: Data, state: Array, spec: int | mjtState) → Data

Sets state in mjx.Data. This is equivalent to mujoco.mj_setState.

Parameters:

• m – model describing the simulation

• d – data for the simulation

• state – a flat array of state values

• spec – int bitmask or mjtState enum specifying which state components to include

Returns:

data with state set to provided values

state_size(m: Model, spec: int | mjtState) → int

Returns the size of a state vector for a given spec.

Parameters:

• m – model describing the simulation

• spec – int bitmask or mjtState enum specifying which state components to include

Returns:

size of the state vector

passive(m: Model, d: Data) → Data

Adds all passive forces.

ray(m: Model, d: Data, pnt: Array, vec: Array, geomgroup: Sequence[int] = (), flg_static: bool = True, bodyexclude: Sequence[int] | int = -1) → Tuple[Array, Array]

Returns the geom id and distance at which a ray intersects with a geom.

Parameters:

• m – MJX model

• d – MJX data

• pnt – ray origin point (3,)

• vec – ray direction (3,)

• geomgroup – group inclusion/exclusion mask, or empty to ignore

• flg_static – if True, allows rays to intersect with static geoms

• bodyexclude – ignore geoms on specified body id or sequence of body ids

Returns:

Distance from ray origin to geom surface (or -1.0 for no intersection) and id of intersected geom (or -1 for no intersection)

render(m: Model, d: Data, ctx: Any) → Data

Render.

get_depth(rc: RenderContextPytree, cam_id: int, depth_data: Array, depth_scale: float) → Array

Extract and normalize depth data for a camera.

Parameters:

• rc – RenderContextPytree.

• cam_id – Camera index to extract.

• depth_data – Raw depth output, shape (total_pixels,) as float32.

• depth_scale – Scale factor for normalizing depth values.

Returns:

Float32 depth array with shape (H, W), clamped to [0, 1].

Raises:

RuntimeError – If Warp is not installed.

get_rgb(rc: RenderContextPytree, cam_id: int, rgb_data: Array) → Array

Unpack uint32 ABGR pixel data into float32 RGB.

Parameters:

• rc – RenderContextPytree.

• cam_id – Camera index to extract.

• rgb_data – Packed render output, shape (total_pixels,) as uint32.

Returns:

Float32 RGB array with shape (H, W, 3), values in [0, 1].

Raises:

RuntimeError – If Warp is not installed.

sensor_acc(m: Model, d: Data) → Data

Compute acceleration/force-dependent sensors values.

sensor_pos(m: Model, d: Data) → Data

Compute position-dependent sensors values.

sensor_vel(m: Model, d: Data) → Data

Compute velocity-dependent sensors values.

camlight(m: Model, d: Data) → Data

Computes camera and light positions and orientations.

com_pos(m: Model, d: Data) → Data

Maps inertias and motion dofs to global frame centered at subtree-CoM.

com_vel(m: Model, d: Data) → Data

Computes cvel, cdof_dot.

crb(m: Model, d: Data) → Data

Runs composite rigid body inertia algorithm.

factor_m(m: Model, d: Data) → Data

Gets factorizaton of inertia-like matrix M, assumed spd.

kinematics(m: Model, d: Data) → Data

Converts position/velocity from generalized coordinates to maximal.

rne(m: Model, d: Data, flg_acc: bool = False) → Data

Computes inverse dynamics using the recursive Newton-Euler algorithm.

flg_acc=False removes inertial term.

rne_postconstraint(m: Model, d: Data) → Data

RNE with complete data: compute cacc, cfrc_ext, cfrc_int.

subtree_vel(m: Model, d: Data) → Data

Subtree linear velocity and angular momentum.

tendon(m: Model, d: Data) → Data

Computes tendon lengths and moments.

tendon_armature(m: Model, d: Data) → Data

Add tendon armature to qM.

tendon_bias(m: Model, d: Data) → Data

Add bias force due to tendon armature.

transmission(m: Model, d: Data) → Data

Computes actuator/transmission lengths and moments.

solve(m: Model, d: Data) → Data

Finds forces that satisfy constraints using conjugate gradient descent.

apply_ft(m: Model, d: Data, force: Array, torque: Array, point: Array, body_id: Array) → Array

Apply Cartesian force and torque.

full_m(m: Model, d: Data) → Array

Reconstitute dense mass matrix from qM.

id2name(m: Model | MjModel, typ: mjtObj, i: int) → str | None

Gets the name of an object with the specified mjtObj type and ids.

See mujoco.id2name for more info.

Parameters:

• m – mujoco.MjModel or mjx.Model

• typ – mujoco.mjtObj type

• i – the id

Returns:

the name string, or None if not found

is_sparse(m: MjModel | Model) → bool

Return True if this model should create sparse mass matrices.

Parameters:

m – a MuJoCo or MJX model

Returns:

True if provided model should create sparse mass matrices

Modern TPUs have specialized hardware for rapidly operating over sparse matrices, whereas GPUs tend to be faster with dense matrices as long as they fit onto the device. As such, the default behavior in MJX (via JacobianType.AUTO) is sparse if nv is >= 60 or MJX detects a TPU as the default backend, otherwise dense.

jac(m: Model, d: Data, point: Array, body_id: Array) → Tuple[Array, Array]

Compute pair of (NV, 3) Jacobians of global point attached to body.

mul_m(m: Model, d: Data, vec: Array) → Array

Multiply vector by inertia matrix.

name2id(m: Model | MjModel, typ: mjtObj, name: str) → int

Gets the id of an object with the specified mjtObj type and name.

See mujoco.mj_name2id for more info.

Parameters:

• m – mujoco.MjModel or mjx.Model

• typ – mujoco.mjtObj type

• name – the name of the object

Returns:

the id, or -1 if not found

xfrc_accumulate(m: Model, d: Data) → Array

Accumulate xfrc_applied into a qfrc.

benchmark(m: MjModel, nstep: int = 1000, batch_size: int = 1024, unroll_steps: int = 1, solver: str = 'newton', iterations: int = 1, ls_iterations: int = 4) → Tuple[float, float, int]

Benchmark a model.

class BiasType

Type of actuator bias.

NONE

no bias

AFFINE

const + kp*length + kv*velocity

MUSCLE

muscle passive force computed by muscle_bias

class CamLightType

Type of camera light.

FIXED

pos and rot fixed in body

TRACK

pos tracks body, rot fixed in global

TRACKCOM

pos tracks subtree com, rot fixed in body

TARGETBODY

pos fixed in body, rot tracks target body

TARGETBODYCOM

pos fixed in body, rot tracks target subtree com

class ConeType

Type of friction cone.

PYRAMIDAL

pyramidal

ELLIPTIC

elliptic

class ConstraintType

Type of constraint.

EQUALITY

equality constraint

LIMIT_JOINT

joint limit

LIMIT_TENDON

tendon limit

CONTACT_FRICTIONLESS

frictionless contact

CONTACT_PYRAMIDAL

frictional contact, pyramidal friction cone

class Contact

Result of collision detection functions.

dist

distance between nearest points; neg: penetration

Type:

jax.Array

pos

position of contact point: midpoint between geoms (3,)

Type:

jax.Array

frame

normal is in [0-2] (9,)

Type:

jax.Array

includemargin

include if dist<includemargin=margin-gap (1,)

Type:

jax.Array

friction

tangent1, 2, spin, roll1, 2 (5,)

Type:

jax.Array

solref

constraint solver reference, normal direction (mjNREF,)

Type:

jax.Array

solreffriction

constraint solver reference, friction directions (mjNREF,)

Type:

jax.Array

solimp

constraint solver impedance (mjNIMP,)

Type:

jax.Array

dim

contact space dimensionality: 1, 3, 4, or 6

Type:

numpy.ndarray

geom1

id of geom 1; deprecated, use geom[0]

Type:

jax.Array

geom2

id of geom 2; deprecated, use geom[1]

Type:

jax.Array

geom

geom ids (2,)

Type:

jax.Array

efc_address

address in efc; -1: not included

Type:

numpy.ndarray

class ConvexMesh

Geom properties for convex meshes.

vert

vertices of the convex mesh

Type:

jax.Array

face

faces of the convex mesh

Type:

jax.Array

face_normal

normal vectors for the faces

Type:

jax.Array

edge

edge indexes for all edges in the convex mesh

Type:

jax.Array

edge_face_normal

indexes for face normals adjacent to edges in edge

Type:

jax.Array

class DataC

C-specific data.

class DataCPP

Minimal Data implementation holding only the pointer.

class DataJAX

JAX-specific data.

class DisableBit

Disable default feature bitflags.

CONSTRAINT

entire constraint solver

EQUALITY

equality constraints

FRICTIONLOSS

joint and tendon frictionloss constraints

LIMIT

joint and tendon limit constraints

CONTACT

contact constraints

SPRING

passive spring forces

DAMPER

passive damper forces

GRAVITY

gravitational forces

CLAMPCTRL

clamp control to specified range

WARMSTART

warmstart constraint solver

ACTUATION

apply actuation forces

REFSAFE

integrator safety: make ref[0]>=2*timestep

SENSOR

sensors

class DynType

Type of actuator dynamics.

NONE

no internal dynamics; ctrl specifies force

INTEGRATOR

integrator: da/dt = u

FILTER

linear filter: da/dt = (u-a) / tau

FILTEREXACT

linear filter: da/dt = (u-a) / tau, with exact integration

MUSCLE

piece-wise linear filter with two time constants

class EnableBit

Enable optional feature bitflags.

INVDISCRETE

discrete-time inverse dynamics

class EqType

Type of equality constraint.

CONNECT

connect two bodies at a point (ball joint)

WELD

fix relative position and orientation of two bodies

JOINT

couple the values of two scalar joints with cubic

TENDON

couple the lengths of two tendons with cubic

class GainType

Type of actuator gain.

FIXED

fixed gain

AFFINE

const + kp*length + kv*velocity

MUSCLE

muscle FLV curve computed by muscle_gain

class GeomType

Type of geometry.

PLANE

plane

HFIELD

height field

SPHERE

sphere

CAPSULE

capsule

ELLIPSOID

ellipsoid

CYLINDER

cylinder

BOX

box

MESH

mesh

SDF

signed distance field

class Impl

Implementation to use.

class IntegratorType

Integrator mode.

EULER

semi-implicit Euler

RK4

4th-order Runge Kutta

IMPLICITFAST

implicit in velocity, no rne derivative

class JacobianType

Type of constraint Jacobian.

DENSE

dense

SPARSE

sparse

AUTO

sparse if nv>60 and device is TPU, dense otherwise

class JointType

Type of degree of freedom.

FREE

global position and orientation (quat) (7,)

BALL

orientation (quat) relative to parent (4,)

SLIDE

sliding distance along body-fixed axis (1,)

HINGE

rotation angle (rad) around body-fixed axis (1,)

class ModelC

CPU-specific model data.

class ModelCPP

Minimal Model implementation holding only the pointer.

class ModelJAX

JAX-specific model data.

class ObjType

Type of object.

UNKNOWN

unknown object type

BODY

body

XBODY

body, used to access regular frame instead of i-frame

GEOM

geom

SITE

site

CAMERA

camera

class Option

Physics options.

class OptionC

C-specific option.

class OptionJAX

JAX-specific option.

class PyTreeNode

Base class for dataclasses that should act like a JAX pytree node.

This base class additionally avoids type checking errors when using PyType.

class SensorType

Type of sensor.

MAGNETOMETER

magnetometer

CAMPROJECTION

camera projection

RANGEFINDER

rangefinder

JOINTPOS

joint position

TENDONPOS

scalar tendon position

ACTUATORPOS

actuator position

BALLQUAT

ball joint orientation

FRAMEPOS

frame position

FRAMEXAXIS

frame x-axis

FRAMEYAXIS

frame y-axis

FRAMEZAXIS

frame z-axis

FRAMEQUAT

frame orientation, represented as quaternion

SUBTREECOM

subtree centor of mass

CLOCK

simulation time

VELOCIMETER

3D linear velocity, in local frame

GYRO

3D angular velocity, in local frame

JOINTVEL

joint velocity

TENDONVEL

scalar tendon velocity

ACTUATORVEL

actuator velocity

BALLANGVEL

ball joint angular velocity

FRAMELINVEL

3D linear velocity

FRAMEANGVEL

3D angular velocity

SUBTREELINVEL

subtree linear velocity

SUBTREEANGMOM

subtree angular momentum

TOUCH

scalar contact normal forces summed over the sensor zone

CONTACT

contacts which occurred during the simulation

ACCELEROMETER

accelerometer

FORCE

force

TORQUE

torque

ACTUATORFRC

scalar actuator force

JOINTACTFRC

scalar actuator force, measured at the joint

TENDONACTFRC

scalar actuator force, measured at the tendon

FRAMELINACC

3D linear acceleration

FRAMEANGACC

3D angular acceleration

class SolverType

Constraint solver algorithm.

CG

Conjugate gradient (primal)

NEWTON

Newton (primal)

class Statistic

Model statistics (in qpos0).

meaninertia

mean diagonal inertia

Type:

jax.Array

meanmass

mean body mass (not used)

Type:

jax.Array

meansize

mean body size (not used)

Type:

jax.Array

extent

spatial extent (not used)

Type:

jax.Array

center

center of model (not used)

Type:

jax.Array

class StatisticWarp

Warp-specific model statistics.

class TrnType

Type of actuator transmission.

JOINT

force on joint

JOINTINPARENT

force on joint, expressed in parent frame

TENDON

force on tendon

SITE

force on site

class WrapType

Type of tendon wrap object.

JOINT

constant moment arm

PULLEY

pulley used to split tendon

SITE

pass through site

SPHERE

wrap around sphere

CYLINDER

wrap around (infinite) cylinder