scenarioAttitudeFeedbackNumba

Overview

Numba reimplementation of scenarioAttitudeFeedback.

The spacecraft dynamics and effectors stay in C++. Every other module in the pipeline – navigation sensor, guidance reference, tracking-error computation, and feedback controller – is reimplemented as a Making Numba Modules.

[spacecraft C++]
     |  scStateOutMsg
[NumbaSimpleNav]          (full Gauss-Markov noise model, optional sun pointing)
     |  attOutMsg / transOutMsg
[NumbaAttTrackingError] <-- [NumbaInertial3D]  attRefOutMsg
     |  attGuidOutMsg
[NumbaMrpFeedback]        (full RW support, controlLawType flag)
     |  cmdTorqueOutMsg
[extForceTorque C++]

The Numba modules are 1-to-1 in capability with the corresponding C/C++ modules:

  • NumbaSimpleNav matches C++ Module: simpleNav: full 18-state Gauss-Markov error model (position, velocity, attitude, rate, sun-pointing, accumulated-DV), crossTrans/crossAtt coupling flags, optional sun-direction output.

  • NumbaInertial3D matches C Module: inertial3D: fixed inertial attitude reference.

  • NumbaAttTrackingError matches C Module: attTrackingError: MRP tracking error, frame-offset correction.

  • NumbaMrpFeedback matches C Module: mrpFeedback: PD/PID feedback with reaction-wheel angular-momentum augmentation and controlLawType flag.

The script is found in the folder basilisk/examples and executed by using:

python3 scenarioAttitudeFeedbackNumba.py
class scenarioAttitudeFeedbackNumba.NumbaAttTrackingError(*args, **kwargs)[source]

Bases: NumbaModel

FSW: MRP attitude tracking error from navigation and reference messages.

Matches C Module: attTrackingError including the frame-offset correction memory.sigma_R0R (body-fixed offset from reference to control-target frame; default zero).

static UpdateStateImpl(attNavInMsgPayload, attRefInMsgPayload, attGuidOutMsgPayload, memory)[source]

Compute the attitude and rate tracking error in the body frame.

class scenarioAttitudeFeedbackNumba.NumbaInertial3D(*args, **kwargs)[source]

Bases: NumbaModel

Guidance: fixed inertial attitude reference with zero rates.

Matches C Module: inertial3D.

static UpdateStateImpl(attRefOutMsgPayload, memory)[source]

Publish a constant inertial attitude reference with zero rates.

class scenarioAttitudeFeedbackNumba.NumbaMrpFeedback(*args, **kwargs)[source]

Bases: NumbaModel

FSW: MRP PD/PID feedback controller with full reaction-wheel support.

Matches C Module: mrpFeedback including:

  • Reaction-wheel angular-momentum augmentation (rwParamsInMsg, rwSpeedsInMsg, optional rwAvailInMsg).

  • memory.controlLawType: 0 = reference-frame gyroscopic (default), else full-body.

  • Integral feedback (memory.Ki > 0) with memory.integralLimit bound.

  • Feed-forward torque via memory.knownTorquePntB_B.

Memory parameters (set directly before InitializeSimulation()):

K proportional gain on MRP error P rate-error feedback gain [N m s] Ki integral gain; negative -> disabled integralLimit wind-up bound [N m] knownTorquePntB_B feed-forward torque [N m, 3-vector] controlLawType 0 = reference-frame gyroscopic; else full-body

static UpdateStateImpl(guidInMsgPayload, vehConfigInMsgPayload, rwSpeedsInMsgPayload, rwSpeedsInMsgIsLinked, rwAvailInMsgPayload, rwAvailInMsgIsLinked, cmdTorqueOutMsgPayload, memory, CurrentSimNanos)[source]

Apply the MRP feedback law and publish the requested body torque.

class scenarioAttitudeFeedbackNumba.NumbaSimpleNav(*args, **kwargs)[source]

Bases: NumbaModel

Navigation sensor with full 18-state Gauss-Markov error model.

Matches C++ Module: simpleNav in capability:

  • 18 error states: position [0:3], velocity [3:6], attitude [6:9], body-rate [9:12], sun-pointing [12:15], accumulated-DV [15:18].

  • memory.PMatrix (18x18 ndarray): process-noise std-deviation matrix.

  • memory.walkBounds (18-vector): random-walk bounds per state.

  • memory.crossTrans / memory.crossAtt (int32): coupling flags.

  • self.RNGSeed: reproducible noise seed (inherited from SysModel, default 0x1badcad1). Set before InitializeSimulation().

  • Optional sun-direction output when sunStateInMsg is subscribed.

Set all memory.* attributes before InitializeSimulation(). Defaults are zero -> noise-free pass-through.

static UpdateStateImpl(scStateInMsgPayload, sunStateInMsgPayload, sunStateInMsgIsLinked, attOutMsgPayload, transOutMsgPayload, CurrentSimNanos, memory, rng)[source]

Propagate the error-state model and publish noisy nav estimates.

scenarioAttitudeFeedbackNumba.run(show_plots, useUnmodeledTorque, useIntGain, useKnownTorque)[source]
Parameters:

  • show_plots (bool) – display plots when True

  • useUnmodeledTorque (bool) – apply a constant external disturbance torque

  • useIntGain (bool) – enable integral feedback in the controller

  • useKnownTorque (bool) – feed-forward the known external torque