Modules:
| Name | Description |
|---|---|
actions |
|
events |
Ground-mode modal-transition events. |
observations |
|
rewards |
|
Classes:
| Name | Description |
|---|---|
FlightActionMapper |
Map batched normalized actions into controller targets, wrench, or RPM. |
FlightMappingConfig |
Mapping-range configuration for normalized action scaling. |
Mixer |
Map batched thrust and moment demands to rotor speed commands. |
AerodynamicEffects |
Compute drag, ground effect, and planar near-wall effect in batches. |
AerodynamicEffectsCfg |
Environment-specific aerodynamic-effect configuration. |
VehicleParams |
Generic multi-rotor vehicle physics. |
RotorDynamics |
Batched motor dynamics model for rotor thrust and torque. |
FlightControlAction |
Body torque control action term. |
FlightControlActionCfg |
See :class: |
DifferentialDriveController |
PX4-style position/heading cascade with differential speed allocation. |
DifferentialDriveActionMapper |
Map normalized ground actions into differential controller targets. |
DifferentialDriveMappingConfig |
Policy-facing scaling for differential-drive controller modes. |
GroundDriveActuatorAdapter |
Batched yaw allocation, side limiting, and actuator command latency. |
GroundDriveActuatorConfig |
Nominal actuator response applied after differential allocation. |
DifferentialDriveControlAction |
Map selectable ground-control actions to drive-joint velocity targets. |
DifferentialDriveControlActionCfg |
Configuration for selectable differential-drive control. |
Functions:
| Name | Description |
|---|---|
randomize_action_term_attr |
Randomize a selected action-term attribute. |
randomize_ground_actuator_params |
Randomize the ground actuator adapter attached to an action term. |
hold_joint_position_target |
Re-assert a fixed joint position target on reset. |
scale_body_inertia |
Apply scalar or per-axis Isaac inertia scales from the default tensor. |
base_roll_pitch |
Return roll and pitch without exposing task-irrelevant absolute yaw. |
ground_velocity_commands |
Return the controllable nonholonomic command channels: forward speed and yaw rate. |
track_lin_vel_x_exp |
Reward forward-speed tracking without treating lateral speed as a command. |
lateral_velocity_l2 |
Penalize nonholonomic lateral slip separately from command tracking. |
Attributes:
| Name | Type | Description |
|---|---|---|
ACTION_KEY_ORDER |
Stable policy-action field order derived from the shared controller schema. |
module-attribute
ACTION_KEY_ORDER = tuple(ControllerState.__annotations__)
Stable policy-action field order derived from the shared controller schema.
Only fields present in the active control mask consume action slices.
FlightActionMapper(control_mode: str, params: VehicleParams = VehicleParams(), mapping: FlightMappingConfig | None = None, control_mask: dict | None = None, num_envs: int = 1, device=None, *, backend: BackendName | None = None, xp: ModuleType | None = None)
Map batched normalized actions into controller targets, wrench, or RPM.
Initialize the batched action mapper and backend namespace.
Methods:
| Name | Description |
|---|---|
randomize |
Randomize. |
action_to_collective_acc |
Action to collective acc. |
collective_acc_to_thrust |
Collective acc to thrust. |
action_to_collective_thrust |
Action to collective thrust. |
decode_action |
Decode normalized batched action into a runtime-neutral control request. |
map_action |
Resolve normalized batched action into simulation rotor RPM commands. |
Attributes:
| Name | Type | Description |
|---|---|---|
requires_controller |
bool
|
Requires controller. |
property
requires_controller: bool
Requires controller.
randomize(env_ids=None, randomization: dict | None = None)
Randomize.
action_to_collective_acc(action: ArrayLike, mixer) -> ArrayLike
Action to collective acc.
Returns:
| Type | Description |
|---|---|
ArrayLike
|
The total collective acceleration array. |
collective_acc_to_thrust(collective_acc: ArrayLike) -> ArrayLike
Collective acc to thrust.
Returns:
| Type | Description |
|---|---|
ArrayLike
|
The total thrust array. |
action_to_collective_thrust(action: ArrayLike, mixer) -> ArrayLike
Action to collective thrust.
Returns:
| Type | Description |
|---|---|
ArrayLike
|
The total thrust array. |
decode_action(action: ArrayLike, mixer, *, state: ControllerState | None = None) -> FlightMapping
Decode normalized batched action into a runtime-neutral control request.
Returns:
| Type | Description |
|---|---|
FlightMapping
|
FlightMapping dictionary containing control requests. |
Raises:
| Type | Description |
|---|---|
ValueError
|
If state is missing when required for control mode. |
map_action(action: ArrayLike, mixer, *, state: ControllerState | None = None, flight_controller: ControllerBase | None = None) -> FlightMapping
Resolve normalized batched action into simulation rotor RPM commands.
Returns:
| Type | Description |
|---|---|
FlightMapping
|
FlightMapping dictionary containing RPM commands and wrench details. |
Raises:
| Type | Description |
|---|---|
ValueError
|
If flight_controller is missing for mapped modes. |
dataclass
FlightMappingConfig(velocity_range: tuple[float, float, float] | None = (3.0, 3.0, 1.5), acceleration_range: tuple[float, float, float] | None = (3.0, 3.0, 2.0), attitude_range: tuple[float, float, float] | None = None, body_rate_range: tuple[float, float, float] | None = (2.0, 2.0, 2.0), rel_rotation: bool = False, rotation_step_len: float | None = None)
Mapping-range configuration for normalized action scaling.
Methods:
| Name | Description |
|---|---|
from_params |
Build mapping ranges from vehicle parameters. |
resolve |
Resolve config overrides into concrete mapping ranges. |
classmethod
from_params(params: VehicleParams) -> FlightMappingConfig
Build mapping ranges from vehicle parameters.
Returns:
| Type | Description |
|---|---|
FlightMappingConfig
|
Instantiated FlightMappingConfig object. |
resolve(params: VehicleParams) -> FlightMappingConfig
Resolve config overrides into concrete mapping ranges.
Returns:
| Type | Description |
|---|---|
FlightMappingConfig
|
Resolved FlightMappingConfig object. |
Mixer(params: VehicleParams = VehicleParams(), num_envs: int = 1, device=None, *, backend: BackendName | None = None, xp: ModuleType | None = None, allocation_mode: Literal['pinv', 'saturation'] = 'pinv')
Map batched thrust and moment demands to rotor speed commands.
The default "pinv" allocator applies direct pseudo-inverse allocation.
"saturation" enables a staged trim strategy: roll/pitch authority is
trimmed first if needed, and yaw authority is allocated from the remaining
rotor-speed margin.
Initialize the mixer matrix and backend namespace.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
|
VehicleParams
|
Vehicle parameters used to build the allocation matrix. |
VehicleParams()
|
|
int
|
Number of parallel environments represented by the batch. |
1
|
|
Optional backend device, mainly used by Torch. |
None
|
|
|
BackendName | None
|
Explicit Array API backend. If omitted, NumPy is used when
|
None
|
|
ModuleType | None
|
Pre-resolved Array API namespace. Overrides |
None
|
|
Literal['pinv', 'saturation']
|
|
'pinv'
|
Methods:
| Name | Description |
|---|---|
randomize |
Randomize mixer parameters for selected environments. |
calculate_rotor_commands |
Allocate |
randomize(env_ids=None, randomization: dict | None = None)
Randomize mixer parameters for selected environments.
calculate_rotor_commands(control: ArrayLike) -> ArrayLike
Allocate [thrust, roll, pitch, yaw] controls to rotor RPM.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
|
ArrayLike
|
Batched control array with shape |
required |
Returns:
| Type | Description |
|---|---|
ArrayLike
|
Batched rotor speed magnitudes with shape |
AerodynamicEffects(params: VehicleParams = VehicleParams(), cfg: AerodynamicEffectsCfg | None = None, num_envs: int = 1, num_rotors: int = 4, device=None, *, backend: BackendName | None = None, xp: ModuleType | None = None)
Compute drag, ground effect, and planar near-wall effect in batches.
Surface geometry is represented in each environment's local world frame. Rotor-force increments, body forces, and force telemetry are represented in the vehicle body frame. The model owns no simulator objects and can therefore be shared by NumPy, Torch, JAX, and CuPy callers.
Initialize coefficients, plane geometry, and telemetry buffers.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
|
VehicleParams
|
Vehicle physical and aerodynamic coefficients. |
VehicleParams()
|
|
AerodynamicEffectsCfg | None
|
Environment-specific switches, wind, and plane geometry. |
None
|
|
int
|
Number of parallel simulation environments. |
1
|
|
int
|
Number of rotors per environment. |
4
|
|
Optional backend device, mainly used by Torch. |
None
|
|
|
BackendName | None
|
Explicit Array API backend name. |
None
|
|
ModuleType | None
|
Pre-resolved Array API namespace. Overrides |
None
|
Methods:
| Name | Description |
|---|---|
randomize |
Sample per-environment aerodynamic state for selected environments. |
set_external_force |
Set one shared or per-environment world-frame external force. |
set_wall_plane |
Set one shared or per-environment wall plane. |
update |
Compute rotor-force increments and root-body force for one simulation step. |
Attributes:
| Name | Type | Description |
|---|---|---|
enabled |
bool
|
Whether any aerodynamic effect is enabled. |
property
enabled: bool
Whether any aerodynamic effect is enabled.
randomize(env_ids: ArrayLike | list[int] | slice | None = None, parameter_ranges: dict[str, tuple[ArrayLike, ArrayLike]] | None = None, effect_probabilities: dict[str, float] | None = None, effect_scale_ranges: dict[str, tuple[ArrayLike, ArrayLike]] | None = None) -> None
Sample per-environment aerodynamic state for selected environments.
Parameter ranges are absolute values rather than multipliers. Effect
probabilities create exact per-episode on/off masks, while scale ranges
multiply the corresponding computed force increment. If every optional
argument is omitted, configuration is read from
params.randomization["aerodynamics"].
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
|
ArrayLike | list[int] | slice | None
|
Environment indices to resample, or all environments. |
None
|
|
dict[str, tuple[ArrayLike, ArrayLike]] | None
|
Absolute ranges for aerodynamic coefficients. |
None
|
|
dict[str, float] | None
|
Bernoulli probability for each effect. |
None
|
|
dict[str, tuple[ArrayLike, ArrayLike]] | None
|
Multiplicative force-scale range per effect. |
None
|
Raises:
| Type | Description |
|---|---|
ValueError
|
If a key, probability, or physical range is invalid. |
set_external_force(external_force_w: ArrayLike) -> None
Set one shared or per-environment world-frame external force.
set_wall_plane(point_w: ArrayLike, normal_w: ArrayLike) -> None
Set one shared or per-environment wall plane.
Moving-surface scenarios can call this before :meth:update to
update the analytic plane pose.
Raises:
| Type | Description |
|---|---|
ValueError
|
If any wall normal is zero, near-zero, or non-finite. |
update(rotor_thrusts: ArrayLike, rotor_positions_w: ArrayLike, rotor_axes_b: ArrayLike, body_velocity_b: ArrayLike, body_rotation_wb: ArrayLike) -> tuple[ArrayLike, ArrayLike]
Compute rotor-force increments and root-body force for one simulation step.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
|
ArrayLike
|
Positive rotor thrust magnitudes, shape |
required |
|
ArrayLike
|
Rotor centers in environment-local world frame,
shape |
required |
|
ArrayLike
|
Rotor positive-thrust axes in vehicle body frame, shape
|
required |
|
ArrayLike
|
Root-body velocity in the vehicle body frame,
shape |
required |
|
ArrayLike
|
Body-to-world rotation matrices, shape
|
required |
Returns:
| Type | Description |
|---|---|
ArrayLike
|
Tuple of per-rotor vehicle-body-frame force increments |
ArrayLike
|
and root-body-frame drag plus prescribed force |
dataclass
AerodynamicEffectsCfg(*, enable_drag: bool = False, enable_external_force: bool = False, enable_ground_effect: bool = False, enable_wall_effect: bool = False, external_force_w: tuple[float, float, float] = (0.0, 0.0, 0.0), ground_height: float = 0.0, max_ground_effect_gain: float = 2.0, wall_plane_point_w: tuple[float, float, float] = (0.0, 0.0, 0.0), wall_plane_normal_w: tuple[float, float, float] = (1.0, 0.0, 0.0), max_surface_distance: float = 5.0)
Environment-specific aerodynamic-effect configuration.
Plane coordinates use each simulator environment's local world frame. Effects default to disabled so existing tasks retain their current physics.
Attributes:
| Name | Type | Description |
|---|---|---|
enable_drag |
bool
|
Apply body-axis linear drag. |
enable_external_force |
bool
|
Apply the configured world-frame external force. |
enable_ground_effect |
bool
|
Augment per-rotor thrust near a horizontal ground plane. |
enable_wall_effect |
bool
|
Apply per-rotor forces near the configured wall plane. |
external_force_w |
tuple[float, float, float]
|
Constant environment-local world-frame external force (N). |
ground_height |
float
|
Height of the analytic horizontal ground plane (m). |
max_ground_effect_gain |
float
|
Maximum ground-effect thrust multiplier used near the model singularity. |
wall_plane_point_w |
tuple[float, float, float]
|
One point on the analytic wall plane (m). |
wall_plane_normal_w |
tuple[float, float, float]
|
Wall-plane unit-normal direction; either side of the plane is valid. |
max_surface_distance |
float
|
Distance beyond which surface effects are exactly disabled (m). |
class-attribute
instance-attribute
enable_drag: bool = False
Apply body-axis linear drag.
class-attribute
instance-attribute
enable_external_force: bool = False
Apply the configured world-frame external force.
class-attribute
instance-attribute
enable_ground_effect: bool = False
Augment per-rotor thrust near a horizontal ground plane.
class-attribute
instance-attribute
enable_wall_effect: bool = False
Apply per-rotor forces near the configured wall plane.
class-attribute
instance-attribute
external_force_w: tuple[float, float, float] = (0.0, 0.0, 0.0)
Constant environment-local world-frame external force (N).
class-attribute
instance-attribute
ground_height: float = 0.0
Height of the analytic horizontal ground plane (m).
class-attribute
instance-attribute
max_ground_effect_gain: float = 2.0
Maximum ground-effect thrust multiplier used near the model singularity.
class-attribute
instance-attribute
wall_plane_point_w: tuple[float, float, float] = (0.0, 0.0, 0.0)
One point on the analytic wall plane (m).
class-attribute
instance-attribute
wall_plane_normal_w: tuple[float, float, float] = (1.0, 0.0, 0.0)
Wall-plane unit-normal direction; either side of the plane is valid.
class-attribute
instance-attribute
max_surface_distance: float = 5.0
Distance beyond which surface effects are exactly disabled (m).
dataclass
VehicleParams(*, rho: float = 1.184, g: float = 9.81, sim_dt: float = 0.01, step_dt: float = 0.01, randomization: VehicleRandomization = VehicleRandomization(), mass: float = 2.1, inertia: list = (lambda: [[0.015, 0.0, 0.0], [0.0, 0.0348, 0.0], [0.0, 0.0, 0.042503]])(), x: float = 0.0952, y: float = 0.114423, h: float = -0.0125, alpha: float = math.radians(8.0), diameter: float = 0.476, drag_coefficients: tuple[float, float, float] = (0.1, 0.1, 0.1), ground_effect_k: float = 0.171, wall_effect_a1: float = 0.21, wall_effect_b1: float = 0.34, wall_effect_a2: float = 0.06, wall_effect_b2: float = 0.54, r_p: float = 0.0775, Ct: float = 0.666, Cq: float = 0.0716, theta0: float = math.radians(14.6), thetaTip: float = math.radians(6.8), theta1: float = math.nan, lock: float = 0.6051, cT: float = math.nan, cM: float = math.nan, spin_dir: tuple[float, float, float, float] = (1.0, 1.0, -1.0, -1.0), tau_f: float = 0.01, tau_m: float = 0.05, max_rpm: float = 10000.0, init_rpm: float = math.nan, rotor_rpm_rate_limit: float = 50000.0, tau_up: float = math.nan, tau_down: float = math.nan, mpc_xy_vel_max: float = 12.0, mpc_z_vel_max_up: float = 3.0, mpc_z_vel_max_dn: float = 1.5, mpc_acc_hor_max: float = 5.0, mpc_acc_up_max: float = 4.0, mpc_acc_down_max: float = 3.0, mpc_man_tilt_max: float = math.radians(35.0), mc_rollrate_max: float = math.radians(220.0), mc_pitchrate_max: float = math.radians(220.0), mc_yawrate_max: float = math.radians(200.0))
Generic multi-rotor vehicle physics.
The base class carries LAV2-compatible defaults for the shared multi-rotor
parameters so that VehicleParams() continues to produce a working
LAV2-flavoured flight configuration without the ground chassis (2.1 kg).
Concrete subclasses add robot-specific locomotion and geometry:
TrackVehicleParams— LAV2: tracked ground locomotion + ducted airframeTiltWheelVehicleParams— ATMO: tilting arms + two side motors whose belt transmissions synchronize four wheel shells
Methods:
| Name | Description |
|---|---|
__post_init__ |
Calculate derived parameters and resolve default sentinels. |
Attributes:
| Name | Type | Description |
|---|---|---|
rho |
float
|
Air density (kg/m³). |
g |
float
|
Gravitational acceleration (m/s²). |
sim_dt |
float
|
Simulation time step (s). |
step_dt |
float
|
Control time step (s). |
randomization |
VehicleRandomization
|
Per-component runtime-randomization configuration. |
mass |
float
|
Total vehicle mass (kg); concrete robot subclasses may override it. |
inertia |
list
|
Vehicle inertia matrix at its center of mass (kg*m^2). |
x |
float
|
Arm length in x — rotor moment-arm for roll (m). |
y |
float
|
Arm length in y — rotor moment-arm for pitch (m). |
h |
float
|
CG height below rotor plane (m, negative = below). |
alpha |
float
|
Duct / rotor cant angle about y-axis (rad — 0 for flat rotors like ATMO). |
diameter |
float
|
Overall vehicle diameter (m). |
drag_coefficients |
tuple[float, float, float]
|
Linear body-axis drag coefficients (N·s/m). |
ground_effect_k |
float
|
Empirical ground-effect distance coefficient for the target multirotor. |
wall_effect_a1 |
float
|
Near-wall lateral-force amplitude fitted by Ding et al. (2023). |
wall_effect_b1 |
float
|
Near-wall lateral-force exponential base fitted by Ding et al. (2023). |
wall_effect_a2 |
float
|
Near-wall vertical-thrust amplitude fitted by Ding et al. (2023). |
wall_effect_b2 |
float
|
Near-wall vertical-thrust exponential base fitted by Ding et al. (2023). |
r_p |
float
|
Propeller radius (m). |
Ct |
float
|
Thrust coefficient (blade-element). |
Cq |
float
|
Torque coefficient (blade-element). |
theta0 |
float
|
Blade root pitch angle (rad). |
thetaTip |
float
|
Blade tip pitch angle (rad). |
theta1 |
float
|
Blade twist (rad), defaults to theta0 - thetaTip. |
lock |
float
|
Lock number. |
cT |
float
|
Thrust coefficient T = cT·rpm² (N / rpm²); defaults from Ct. |
cM |
float
|
Reaction-torque coefficient M = cM·rpm² (N·m / rpm²); defaults from Cq. |
spin_dir |
tuple[float, float, float, float]
|
Rotor spin direction signs; + means clockwise, - means counter-clockwise. |
tau_f |
float
|
Motor filter time constant (s). |
tau_m |
float
|
Legacy motor time constant (s). |
max_rpm |
float
|
Maximum rotor speed (rpm — LAV2 default). |
init_rpm |
float
|
Initial rotor speed — defaults to 0.2·max_rpm. |
rotor_rpm_rate_limit |
float
|
Rotor-speed rate-of-change limit (rpm/s). |
tau_up |
float
|
Motor spin-up time constant (s), defaults to tau_m. |
tau_down |
float
|
Motor spin-down time constant (s), defaults to tau_m. |
mpc_xy_vel_max |
float
|
Maximum horizontal velocity setpoint magnitude in m/s (0 - 20). |
mpc_z_vel_max_up |
float
|
Maximum upward velocity setpoint in m/s (0.5 - 8). |
mpc_z_vel_max_dn |
float
|
Maximum downward velocity setpoint magnitude in m/s (0.5 - 4). |
mpc_acc_hor_max |
float
|
Maximum horizontal acceleration setpoint magnitude in m/s^2 (2 - 15). |
mpc_acc_up_max |
float
|
Maximum upward acceleration setpoint in m/s^2 (2 - 15). |
mpc_acc_down_max |
float
|
Maximum downward acceleration setpoint magnitude in m/s^2 (2 - 15). |
mpc_man_tilt_max |
float
|
Maximum manual tilt angle in radians (deg: 0 - 90). |
mc_rollrate_max |
float
|
Maximum roll-rate setpoint in rad/s (deg/s: 0 - 1800). |
mc_pitchrate_max |
float
|
Maximum pitch-rate setpoint in rad/s (deg/s: 0 - 1800). |
mc_yawrate_max |
float
|
Maximum yaw-rate setpoint in rad/s (deg/s: 0 - 1800). |
class-attribute
instance-attribute
rho: float = 1.184
Air density (kg/m³).
class-attribute
instance-attribute
g: float = 9.81
Gravitational acceleration (m/s²).
class-attribute
instance-attribute
sim_dt: float = 0.01
Simulation time step (s).
class-attribute
instance-attribute
step_dt: float = 0.01
Control time step (s).
class-attribute
instance-attribute
randomization: VehicleRandomization = field(default_factory=VehicleRandomization)
Per-component runtime-randomization configuration.
class-attribute
instance-attribute
mass: float = 2.1
Total vehicle mass (kg); concrete robot subclasses may override it.
class-attribute
instance-attribute
inertia: list = field(default_factory=lambda: [[0.015, 0.0, 0.0], [0.0, 0.0348, 0.0], [0.0, 0.0, 0.042503]])
Vehicle inertia matrix at its center of mass (kg*m^2).
class-attribute
instance-attribute
x: float = 0.0952
Arm length in x — rotor moment-arm for roll (m).
class-attribute
instance-attribute
y: float = 0.114423
Arm length in y — rotor moment-arm for pitch (m).
class-attribute
instance-attribute
h: float = -0.0125
CG height below rotor plane (m, negative = below).
class-attribute
instance-attribute
alpha: float = math.radians(8.0)
Duct / rotor cant angle about y-axis (rad — 0 for flat rotors like ATMO).
class-attribute
instance-attribute
diameter: float = 0.476
Overall vehicle diameter (m).
class-attribute
instance-attribute
drag_coefficients: tuple[float, float, float] = (0.1, 0.1, 0.1)
Linear body-axis drag coefficients (N·s/m).
class-attribute
instance-attribute
ground_effect_k: float = 0.171
Empirical ground-effect distance coefficient for the target multirotor.
class-attribute
instance-attribute
wall_effect_a1: float = 0.21
Near-wall lateral-force amplitude fitted by Ding et al. (2023).
class-attribute
instance-attribute
wall_effect_b1: float = 0.34
Near-wall lateral-force exponential base fitted by Ding et al. (2023).
class-attribute
instance-attribute
wall_effect_a2: float = 0.06
Near-wall vertical-thrust amplitude fitted by Ding et al. (2023).
class-attribute
instance-attribute
wall_effect_b2: float = 0.54
Near-wall vertical-thrust exponential base fitted by Ding et al. (2023).
class-attribute
instance-attribute
r_p: float = 0.0775
Propeller radius (m).
class-attribute
instance-attribute
Ct: float = 0.666
Thrust coefficient (blade-element).
class-attribute
instance-attribute
Cq: float = 0.0716
Torque coefficient (blade-element).
class-attribute
instance-attribute
theta0: float = math.radians(14.6)
Blade root pitch angle (rad).
class-attribute
instance-attribute
thetaTip: float = math.radians(6.8)
Blade tip pitch angle (rad).
class-attribute
instance-attribute
theta1: float = math.nan
Blade twist (rad), defaults to theta0 - thetaTip.
class-attribute
instance-attribute
lock: float = 0.6051
Lock number.
class-attribute
instance-attribute
cT: float = math.nan
Thrust coefficient T = cT·rpm² (N / rpm²); defaults from Ct.
class-attribute
instance-attribute
cM: float = math.nan
Reaction-torque coefficient M = cM·rpm² (N·m / rpm²); defaults from Cq.
class-attribute
instance-attribute
spin_dir: tuple[float, float, float, float] = (1.0, 1.0, -1.0, -1.0)
Rotor spin direction signs; + means clockwise, - means counter-clockwise.
LAV2 defaults to rotors 0/1 clockwise and 2/3 counter-clockwise, so the
signs are (+,+,-,-).
class-attribute
instance-attribute
tau_f: float = 0.01
Motor filter time constant (s).
class-attribute
instance-attribute
tau_m: float = 0.05
Legacy motor time constant (s).
class-attribute
instance-attribute
max_rpm: float = 10000.0
Maximum rotor speed (rpm — LAV2 default).
class-attribute
instance-attribute
init_rpm: float = math.nan
Initial rotor speed — defaults to 0.2·max_rpm.
class-attribute
instance-attribute
rotor_rpm_rate_limit: float = 50000.0
Rotor-speed rate-of-change limit (rpm/s).
class-attribute
instance-attribute
tau_up: float = math.nan
Motor spin-up time constant (s), defaults to tau_m.
class-attribute
instance-attribute
tau_down: float = math.nan
Motor spin-down time constant (s), defaults to tau_m.
class-attribute
instance-attribute
mpc_xy_vel_max: float = 12.0
Maximum horizontal velocity setpoint magnitude in m/s (0 - 20).
class-attribute
instance-attribute
mpc_z_vel_max_up: float = 3.0
Maximum upward velocity setpoint in m/s (0.5 - 8).
class-attribute
instance-attribute
mpc_z_vel_max_dn: float = 1.5
Maximum downward velocity setpoint magnitude in m/s (0.5 - 4).
class-attribute
instance-attribute
mpc_acc_hor_max: float = 5.0
Maximum horizontal acceleration setpoint magnitude in m/s^2 (2 - 15).
class-attribute
instance-attribute
mpc_acc_up_max: float = 4.0
Maximum upward acceleration setpoint in m/s^2 (2 - 15).
class-attribute
instance-attribute
mpc_acc_down_max: float = 3.0
Maximum downward acceleration setpoint magnitude in m/s^2 (2 - 15).
class-attribute
instance-attribute
mpc_man_tilt_max: float = math.radians(35.0)
Maximum manual tilt angle in radians (deg: 0 - 90).
class-attribute
instance-attribute
mc_rollrate_max: float = math.radians(220.0)
Maximum roll-rate setpoint in rad/s (deg/s: 0 - 1800).
class-attribute
instance-attribute
mc_pitchrate_max: float = math.radians(220.0)
Maximum pitch-rate setpoint in rad/s (deg/s: 0 - 1800).
class-attribute
instance-attribute
mc_yawrate_max: float = math.radians(200.0)
Maximum yaw-rate setpoint in rad/s (deg/s: 0 - 1800).
__post_init__()
Calculate derived parameters and resolve default sentinels.
Raises:
| Type | Description |
|---|---|
ValueError
|
If spin_dir does not contain four signs. |
RotorDynamics(params: VehicleParams = VehicleParams(), num_envs: int = 1, device=None, *, backend: BackendName | None = None, xp: ModuleType | None = None)
Bases: DynamicsBase
Batched motor dynamics model for rotor thrust and torque.
This backend-agnostic Array API implementation keeps
the same low-pass motor response, RPM rate limiting, quadratic thrust/torque
coefficients, and motor-direction torque signs while operating on
(num_envs, 4) batches.
Initialize rotor dynamics and backend arrays.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
|
VehicleParams
|
Vehicle parameters used by the motor model. |
VehicleParams()
|
|
int
|
Number of parallel environments represented by the batch. |
1
|
|
Optional backend device, mainly used by Torch. |
None
|
|
|
BackendName | None
|
Explicit Array API backend name. |
None
|
|
ModuleType | None
|
Pre-resolved Array API namespace. Overrides |
None
|
Methods:
| Name | Description |
|---|---|
randomize |
Randomize motor parameters for selected environments. |
update |
Update motor state and return thrust/torque outputs. |
reset |
Reset selected environments to initial rotor speeds. |
Attributes:
| Name | Type | Description |
|---|---|---|
params |
VehicleParams
|
Vehicle parameters for the dynamics model. |
instance-attribute
params: VehicleParams = params
Vehicle parameters for the dynamics model.
randomize(env_ids=None, randomization: dict[str, tuple[float, float]] | None = None)
Randomize motor parameters for selected environments.
update(commands: ArrayLike) -> ArrayLike
Update motor state and return thrust/torque outputs.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
|
ArrayLike
|
Desired rotor speeds with shape |
required |
Returns:
| Type | Description |
|---|---|
ArrayLike
|
Batched thrusts followed by torques with shape |
reset(env_ids=None)
Reset selected environments to initial rotor speeds.
FlightControlAction(cfg: FlightControlActionCfg, env: ManagerBasedRLEnv)
Bases: ActionTerm
Body torque control action term.
This action term applies a wrench to the drone body frame based on action commands
Attributes:
| Name | Type | Description |
|---|---|---|
cfg |
FlightControlActionCfg
|
The configuration of the action term. |
aerodynamics |
AerodynamicEffects
|
Shared aerodynamic model and force telemetry. |
IO_descriptor |
GenericActionIODescriptor
|
The IO descriptor of the LAV2 flight control action term. |
instance-attribute
cfg: FlightControlActionCfg = cfg
The configuration of the action term.
property
aerodynamics: AerodynamicEffects
Shared aerodynamic model and force telemetry.
property
IO_descriptor: GenericActionIODescriptor
The IO descriptor of the LAV2 flight control action term.
Bases: ActionTermCfg
See :class:FlightControlAction for more details.
Attributes:
| Name | Type | Description |
|---|---|---|
class_type |
type[ActionTerm]
|
Class of the action term. |
asset_name |
str
|
Name of the asset in the environment for which the commands are generated. |
rotor_body_names |
tuple[str, ...]
|
Ordered rotor body names matching controller/mixer rotor order. |
root_body_name |
str | None
|
Body receiving drag and prescribed force; defaults to the articulation root. |
params |
VehicleParams
|
Vehicle parameters. |
gains |
dict | None
|
Optional gains configuration for the flight controller. |
limits |
dict | None
|
Optional command-limit configuration passed directly to the flight controller. |
controller |
str
|
Controller class name. Options: 'FlightController' (PID), 'GeoControl'. |
control_mode |
str
|
Control mode, options: cmd_motor_thrusts, cmd_motor_speeds, cmd_ctbm, cmd_ctbr, cmd_ctatt, cmd_ctatt_euler, cmd_ctatt_quat, cmd_ctatt_rotmat, cmd_ctatt_rotvec, cmd_acc, cmd_vel, cmd_vel_ffacc, cmd_pos. |
mapping |
FlightMappingConfig
|
Action-mapping range overrides resolved against |
aerodynamics |
AerodynamicEffectsCfg
|
Optional drag, external-force, ground-effect, and wall configuration. |
class-attribute
instance-attribute
class_type: type[ActionTerm] = FlightControlAction
Class of the action term.
class-attribute
instance-attribute
asset_name: str = 'robot'
Name of the asset in the environment for which the commands are generated.
class-attribute
instance-attribute
rotor_body_names: tuple[str, ...] = ('rotor0', 'rotor1', 'rotor2', 'rotor3')
Ordered rotor body names matching controller/mixer rotor order.
class-attribute
instance-attribute
root_body_name: str | None = None
Body receiving drag and prescribed force; defaults to the articulation root.
class-attribute
instance-attribute
params: VehicleParams = VehicleParams()
Vehicle parameters.
class-attribute
instance-attribute
gains: dict | None = None
Optional gains configuration for the flight controller.
class-attribute
instance-attribute
limits: dict | None = None
Optional command-limit configuration passed directly to the flight controller.
class-attribute
instance-attribute
controller: str = 'FlightController'
Controller class name. Options: 'FlightController' (PID), 'GeoControl'.
class-attribute
instance-attribute
control_mode: str = 'cmd_motor_speeds'
Control mode, options: cmd_motor_thrusts, cmd_motor_speeds, cmd_ctbm, cmd_ctbr, cmd_ctatt, cmd_ctatt_euler, cmd_ctatt_quat, cmd_ctatt_rotmat, cmd_ctatt_rotvec, cmd_acc, cmd_vel, cmd_vel_ffacc, cmd_pos.
rotorpy ref: - https://github.com/spencerfolk/rotorpy/blob/main/rotorpy/learning/quadrotor_environments.py#L319C1-L350C12 - https://github.com/spencerfolk/rotorpy/blob/main/rotorpy/vehicles/multirotor.py#L997
TODO: also align control modes with px4, see: https://docs.px4.io/v1.14/en/flight_modes/offboard (version is unclear ATM)
class-attribute
instance-attribute
mapping: FlightMappingConfig = FlightMappingConfig()
Action-mapping range overrides resolved against params.
The default mapping is an intentional policy-facing behavior rather than a direct reflection of the parameter-derived controller ranges.
class-attribute
instance-attribute
aerodynamics: AerodynamicEffectsCfg = AerodynamicEffectsCfg()
Optional drag, external-force, ground-effect, and wall configuration.
DifferentialDriveController(params: VehicleParams, control_mask: dict | None = None, gains: dict | None = None, limits: dict | None = None, num_envs: int = 1, device=None, *, backend: BackendName | None = None, xp: ModuleType | None = None)
Bases: ControllerBase
PX4-style position/heading cascade with differential speed allocation.
The output is [left, right] drive-shaft angular velocity in rad/s. The
controller is independent of wheel count: a platform adapter commands one
physical drive joint per side, while backend mechanical constraints
synchronize any follower road wheels or wheel shells.
Initialize the batched differential-drive controller.
Methods:
| Name | Description |
|---|---|
randomize |
Randomize gains and limits for selected environments. |
reset |
Reset all PID state for selected environments. |
update |
Compute left/right drive-shaft angular-velocity targets in rad/s. |
Attributes:
| Name | Type | Description |
|---|---|---|
params |
VehicleParams
|
Vehicle parameters for the controller. |
instance-attribute
params: VehicleParams = params
Vehicle parameters for the controller.
randomize(env_ids=None, gains: dict[str, Any] | None = None, limits: dict[str, Any] | None = None)
Randomize gains and limits for selected environments.
reset(env_ids=None)
Reset all PID state for selected environments.
update(target: ControllerTarget, state: ControllerState) -> ArrayLike
Compute left/right drive-shaft angular-velocity targets in rad/s.
Returns:
| Type | Description |
|---|---|
ArrayLike
|
The computed drive-shaft angular velocity array. |
DifferentialDriveActionMapper(control_mode: str, params: VehicleParams, mapping: DifferentialDriveMappingConfig | None = None, control_mask: dict | None = None, num_envs: int = 1, device=None, *, backend: BackendName | None = None, xp: ModuleType | None = None)
Map normalized ground actions into differential controller targets.
Initialize the ground mapper and stable action-field indices.
Raises:
| Type | Description |
|---|---|
ValueError
|
If mapping ranges or control mode is invalid. |
Methods:
| Name | Description |
|---|---|
decode_action |
Decode a policy action into side speed or a controller target. |
map_action |
Resolve an action into left/right output-shaft speed targets. |
Attributes:
| Name | Type | Description |
|---|---|---|
requires_controller |
bool
|
Whether this mode passes through the differential controller. |
property
requires_controller: bool
Whether this mode passes through the differential controller.
decode_action(action: ArrayLike, *, state: ControllerState | None = None) -> DifferentialDriveMapping
Decode a policy action into side speed or a controller target.
Returns:
| Type | Description |
|---|---|
DifferentialDriveMapping
|
DifferentialDriveMapping dictionary containing decoded action. |
Raises:
| Type | Description |
|---|---|
ValueError
|
If state is missing when required. |
map_action(action: ArrayLike, *, state: ControllerState | None = None, controller: ControllerBase | None = None) -> DifferentialDriveMapping
Resolve an action into left/right output-shaft speed targets.
Returns:
| Type | Description |
|---|---|
DifferentialDriveMapping
|
DifferentialDriveMapping dictionary containing side speed targets. |
Raises:
| Type | Description |
|---|---|
ValueError
|
If controller is missing when required. |
dataclass
DifferentialDriveMappingConfig(forward_speed_range: float | None = None, yaw_rate_range: float | None = None, heading_range: float = pi, side_speed_range: float | None = None)
Policy-facing scaling for differential-drive controller modes.
Methods:
| Name | Description |
|---|---|
from_params |
Build mapping ranges from track or wheel platform parameters. |
resolve |
Resolve explicit overrides against platform-derived defaults. |
classmethod
from_params(params: VehicleParams) -> DifferentialDriveMappingConfig
Build mapping ranges from track or wheel platform parameters.
Returns:
| Type | Description |
|---|---|
DifferentialDriveMappingConfig
|
Instantiated DifferentialDriveMappingConfig object. |
Raises:
| Type | Description |
|---|---|
TypeError
|
If params lacks track or wheel attributes. |
resolve(params: VehicleParams) -> DifferentialDriveMappingConfig
Resolve explicit overrides against platform-derived defaults.
Returns:
| Type | Description |
|---|---|
DifferentialDriveMappingConfig
|
Resolved DifferentialDriveMappingConfig object. |
GroundDriveActuatorAdapter(*, physics_dt: float, side_speed_limit: float, config: GroundDriveActuatorConfig, num_envs: int = 1, device=None, backend: BackendName | None = None, xp: ModuleType | None = None)
Batched yaw allocation, side limiting, and actuator command latency.
Build an Array API adapter for one fixed-rate physics loop.
Methods:
| Name | Description |
|---|---|
randomize |
Randomize selected actuator parameters as nominal-relative scales. |
reset |
Clear actuator command state for selected environments. |
update |
Return the physical-side target for the current physics step. |
Attributes:
| Name | Type | Description |
|---|---|---|
side_gain |
ArrayLike
|
Current left/right gain samples. |
property
side_gain: ArrayLike
Current left/right gain samples.
randomize(env_ids=None, randomization: dict[str, tuple[float, float] | float] | None = None) -> None
Randomize selected actuator parameters as nominal-relative scales.
reset(env_ids=None) -> None
Clear actuator command state for selected environments.
update(targets: ArrayLike, *, apply_yaw_compensation: bool = True) -> ArrayLike
Return the physical-side target for the current physics step.
dataclass
GroundDriveActuatorConfig(yaw_allocation_scale: float = 1.0, left_gain: float = 1.0, right_gain: float = 1.0, left_delay_s: float = 0.0, right_delay_s: float = 0.0)
Nominal actuator response applied after differential allocation.
DifferentialDriveControlAction(cfg: DifferentialDriveControlActionCfg, env: ManagerBasedRLEnv)
Bases: ActionTerm
Map selectable ground-control actions to drive-joint velocity targets.
Attributes:
| Name | Type | Description |
|---|---|---|
ideal_side_speed_targets |
Tensor
|
Controller output before Isaac-specific allocation and latency. |
applied_side_speed_targets |
Tensor
|
Side targets after allocation/gain and before actuator latency. |
sampled_actuator_parameters |
dict[str, Tensor]
|
Expose per-environment samples for episode diagnostics. |
IO_descriptor |
GenericActionIODescriptor
|
Describe policy channels for the selected ground-control mode. |
property
ideal_side_speed_targets: Tensor
Controller output before Isaac-specific allocation and latency.
property
applied_side_speed_targets: Tensor
Side targets after allocation/gain and before actuator latency.
property
sampled_actuator_parameters: dict[str, Tensor]
Expose per-environment samples for episode diagnostics.
property
IO_descriptor: GenericActionIODescriptor
Describe policy channels for the selected ground-control mode.
Bases: ActionTermCfg
Configuration for selectable differential-drive control.
randomize_action_term_attr(env: ManagerBasedEnv, env_ids: Tensor | None, action_term_name: str = 'flight_control_action', attr_name: str = '_flight_controller', **kwargs) -> None
Randomize a selected action-term attribute.
randomize_ground_actuator_params(env: ManagerBasedRLEnv, env_ids, action_term_name: str = 'ground_control_action', actuator_attr: str = '_actuator_adapter', param_range: dict[str, tuple[float, float] | float] | None = None) -> None
Randomize the ground actuator adapter attached to an action term.
hold_joint_position_target(env: ManagerBasedRLEnv, env_ids, target_pos: float, joint_names: list[str], asset_cfg: SceneEntityCfg = SceneEntityCfg('robot')) -> None
Re-assert a fixed joint position target on reset.
ArticulationCfg.InitialStateCfg.joint_pos sets the initial joint state,
not the actuator target. For joints driven by a stiff position actuator that
are not policy-controlled — e.g. ATMO's base_to_arm[l,r] locked at 90 deg
for the ground-mode modal transition — the drive target defaults to 0 and the
actuator yanks the joint out of its init pose (arms snap back to the 0-deg
flight pose, dropping the chassis onto the ground). This event writes the
position target back to target_pos each reset so the actuator holds the
configured pose for the episode; the target persists across steps because the
action manager never commands these joints.
scale_body_inertia(env: ManagerBasedRLEnv, env_ids, scale: float | tuple[float, float] | tuple[tuple[float, float], tuple[float, float], tuple[float, float]], asset_cfg: SceneEntityCfg = SceneEntityCfg('robot')) -> None
Apply scalar or per-axis Isaac inertia scales from the default tensor.
base_roll_pitch(env: ManagerBasedRLEnv, asset_cfg: SceneEntityCfg = SceneEntityCfg('robot')) -> torch.Tensor
Return roll and pitch without exposing task-irrelevant absolute yaw.
ground_velocity_commands(env: ManagerBasedRLEnv, command_name: str) -> torch.Tensor
Return the controllable nonholonomic command channels: forward speed and yaw rate.
track_lin_vel_x_exp(env: ManagerBasedRLEnv, std: float, command_name: str, asset_cfg: SceneEntityCfg = SceneEntityCfg('robot')) -> torch.Tensor
Reward forward-speed tracking without treating lateral speed as a command.
lateral_velocity_l2(env: ManagerBasedRLEnv, asset_cfg: SceneEntityCfg = SceneEntityCfg('robot')) -> torch.Tensor
Penalize nonholonomic lateral slip separately from command tracking.