Fault Environment Example

This tutorial demonstrates how to configure and use a simple BSK-RL environment to model faults in a system with four reaction wheels (RWs).

Load Modules

[1]:
import numpy as np
from typing import Iterable

from Basilisk.architecture import bskLogging
from Basilisk.utilities import macros, orbitalMotion, simIncludeRW
from Basilisk.simulation import reactionWheelStateEffector
from Basilisk.fswAlgorithms import rwNullSpace
from Basilisk.architecture import messaging
from bsk_rl import SatelliteTasking, act, data, obs, scene, sats
from bsk_rl.sim import dyn, fsw
from bsk_rl.utils.orbital import random_orbit, random_unit_vector
from bsk_rl.utils.functional import default_args

bskLogging.setDefaultLogLevel(bskLogging.BSK_WARNING)

Making Faults Cases

Creating a fault base class and defining individual fault types enables modeling multiple kinds of faults within a single satellite. In this example, a power draw limit is applied to RWs, causing it to operate at reduced speed compared to nominal conditions. By default, while a torque limit is enforced, there are no restrictions on power draw. time is used to define the time at which the fault occurs, reducedLimit specifies the power draw limit in watts, and wheel_Idx indicates which RW is affected by the fault. It can be set to a value from 1 to 4, or to all to apply the fault to every RW.

[2]:
class FaultObject:
    def __init__(self, name, time, verbose=True, **kwargs):
        self.name = name
        self.time = time
        self.verbose = verbose
        self.message = None
        self.message_printed = False

    def execute(self, satellite):
        raise NotImplementedError(
            f"{self.name} does not have a custom execute function!"
        )

    def print_message(self, message, satellite):
        if not self.message_printed:
            satellite.logger.info(message)
            self.message_printed = True

    def addFaultToSimulation(self, satellite, listIdx):
        self.uniqueFaultIdx = listIdx  # Index in the faultList array.
        satellite.simulator.createNewEvent(
            f"add{self.name}Fault",
            satellite.dynamics.dyn_rate,
            eventActive=True,
            conditionTime=self.time,
            actionList=[
                f"self.faultList[{self.uniqueFaultIdx}].execute({satellite._satellite_command})",
                f"self.faultList[{self.uniqueFaultIdx}].print({satellite._satellite_command})",
            ],
        )


class RwPowerFault(FaultObject):
    def __init__(self, name, time, reducedLimit, wheelIdx):
        super().__init__(name, time)
        self.reducedLimit = reducedLimit

        if isinstance(wheelIdx, float):
            # int needed around wheelIdx because np.random.choice doesn't return
            # a type int, and the index will not register in execute without it.
            self.wheelIdx = int(wheelIdx)
        elif isinstance(wheelIdx, int) or wheelIdx == "all":
            # option to trigger the fault in all wheels reflecting a larger power issue
            self.wheelIdx = wheelIdx
        else:
            raise ValueError(
                "Fault parameter 'wheelIdx' must either be a number corresponding to a reaction wheel or the string 'all'"
            )

    def execute(self, satellite):
        dynModels = satellite.dynamics
        if self.wheelIdx == 1:
            dynModels.rwFactory.rwList["RW1"].P_max = self.reducedLimit
        elif self.wheelIdx == 2:
            dynModels.rwFactory.rwList["RW2"].P_max = self.reducedLimit
        elif self.wheelIdx == 3:
            dynModels.rwFactory.rwList["RW3"].P_max = self.reducedLimit
        elif self.wheelIdx == 4:
            dynModels.rwFactory.rwList["RW4"].P_max = self.reducedLimit
        elif self.wheelIdx == "all":
            # option to trigger the fault in all wheels (not supported for all fault types)
            dynModels.rwFactory.rwList["RW1"].P_max = self.reducedLimit
            dynModels.rwFactory.rwList["RW2"].P_max = self.reducedLimit
            dynModels.rwFactory.rwList["RW3"].P_max = self.reducedLimit
            dynModels.rwFactory.rwList["RW4"].P_max = self.reducedLimit

    def print(self, satellite):
        if self.wheelIdx == "all":
            self.message = f"RW Power Fault: all RW's power limit reduced to {self.reducedLimit} Watts at {self.time * macros.NANO2MIN} minutes!"
        else:
            self.message = f"RW Power Fault: RW{self.wheelIdx}'s power limit reduced to {self.reducedLimit} Watts at {self.time * macros.NANO2MIN} minutes!"
        super().print_message(self.message, satellite)

Configure the Simulation Models

  • Dynamics model: FullFeaturedDynModel is used as the base class, and setup_reaction_wheel_dyn_effector is overridden to support four RWs. Two additional properties are added: the angle between the Sun and the solar panel, and the speed fraction of each RW.

[3]:
class CustomDynModel(dyn.FullFeaturedDynModel):
    @property
    def solar_angle_norm(self) -> float:
        sun_vec_N = (
            self.world.gravFactory.spiceObject.planetStateOutMsgs[self.world.sun_index]
            .read()
            .PositionVector
        )
        sun_vec_N_hat = sun_vec_N / np.linalg.norm(sun_vec_N)
        solar_panel_vec_B = np.array([0, 0, -1])
        mat = np.transpose(self.BN)
        solar_panel_vec_N = np.matmul(mat, solar_panel_vec_B)
        error_angle = np.arccos(np.dot(solar_panel_vec_N, sun_vec_N_hat))

        return error_angle / np.pi

    @property
    def wheel_speeds_frac(self):
        rw_speed = self.wheel_speeds
        return rw_speed[0:4] / (self.maxWheelSpeed * macros.rpm2radsec)

    @default_args(
        wheelSpeeds=lambda: np.random.uniform(-1500, 1500, 4),
        maxWheelSpeed=np.inf,
        u_max=0.200,
    )
    def setup_reaction_wheel_dyn_effector(
        self,
        wheelSpeeds: Iterable[float],
        maxWheelSpeed: float,
        u_max: float,
        priority: int = 997,
        **kwargs,
    ) -> None:
        """Set the RW state effector parameters.

        Args:
            wheelSpeeds: Initial speeds of each wheel [RPM]
            maxWheelSpeed: Failure speed for wheels [RPM]
            u_max: Torque producible by wheel [N*m]
            priority: Model priority.
            kwargs: Ignored
        """

        def balancedHR16Triad(
            useRandom=False, randomBounds=(-400, 400), wheelSpeeds=[500, 500, 500, 500]
        ):
            """Create a set of three HR16 reaction wheels.

            Args:
                useRandom: Use random values for wheel speeds.
                randomBounds: Bounds for random wheel speeds.
                wheelSpeeds: Fixed wheel speeds.

            Returns:
                tuple:
                    * **rwStateEffector**: Reaction wheel state effector instance.
                    * **rwFactory**: Factory containing defined reaction wheels.
                    * **wheelSpeeds**: Wheel speeds.
            """
            rwFactory = simIncludeRW.rwFactory()

            if useRandom:
                wheelSpeeds = np.random.uniform(randomBounds[0], randomBounds[1], 4)
            c = 3 ** (-0.5)
            rwFactory.create(
                "Honeywell_HR16",
                [1, 0, 0],
                maxMomentum=50.0,
                Omega=wheelSpeeds[0],
            )
            rwFactory.create(
                "Honeywell_HR16",
                [0, 1, 0],
                maxMomentum=50.0,
                Omega=wheelSpeeds[1],
            )
            rwFactory.create(
                "Honeywell_HR16",
                [0, 0, 1],
                maxMomentum=50.0,
                Omega=wheelSpeeds[2],
            )
            rwFactory.create(
                "Honeywell_HR16",
                [c, c, c],
                maxMomentum=50.0,
                Omega=wheelSpeeds[3],
            )

            rwStateEffector = reactionWheelStateEffector.ReactionWheelStateEffector()

            return rwStateEffector, rwFactory, wheelSpeeds

        self.maxWheelSpeed = maxWheelSpeed
        self.rwStateEffector, self.rwFactory, _ = balancedHR16Triad(
            useRandom=False,
            wheelSpeeds=wheelSpeeds,
        )
        for RW in self.rwFactory.rwList.values():
            RW.u_max = u_max
        self.rwStateEffector.ModelTag = "ReactionWheels"
        self.rwFactory.addToSpacecraft(
            self.scObject.ModelTag, self.rwStateEffector, self.scObject
        )
        self.simulator.AddModelToTask(
            self.task_name, self.rwStateEffector, ModelPriority=priority
        )

        self.Gs = np.array(
            [
                [1, 0, 0, 1 / np.sqrt(3)],  # RW1 and RW4 x-components
                [0, 1, 0, 1 / np.sqrt(3)],  # RW2 and RW4 y-components
                [0, 0, 1, 1 / np.sqrt(3)],  # RW3 and RW4 z-components
            ]
        )
  • Flight software model: A custom flight software model is defined to support four RWs. It is based on the SteeringImagerFSWModel, with the main modification being the addition of the rwNullSpace module. Due to the redundancy of having four RWs, there are infinitely many solutions for mapping the required body control torque to individual RW torques. To address this, once the control torque is computed, the RW null space is used to decelerate the wheels without applying additional torque to the spacecraft.

[4]:
class CustomSteeringImagerFSWModel(fsw.SteeringImagerFSWModel):
    def __init__(self, *args, **kwargs) -> None:
        """Convenience type that combines the imaging FSW model with MRP steering for four reaction wheels."""
        super().__init__(*args, **kwargs)

    def _set_config_msgs(self) -> None:
        super()._set_config_msgs()
        self._set_rw_constellation_msg()

    def _set_rw_constellation_msg(self) -> None:
        """Set the reaction wheel constellation message."""
        rwConstellationConfig = messaging.RWConstellationMsgPayload()
        rwConstellationConfig.numRW = self.dynamics.rwFactory.getNumOfDevices()
        rwConfigElementList = []
        for i in range(4):
            rwConfigElementMsg = messaging.RWConfigElementMsgPayload()
            rwConfigElementMsg.gsHat_B = self.dynamics.Gs[:, i]
            rwConfigElementMsg.Js = self.dynamics.rwFactory.rwList[f"RW{i + 1}"].Js
            rwConfigElementMsg.uMax = self.dynamics.rwFactory.rwList[f"RW{i + 1}"].u_max
            rwConfigElementList.append(rwConfigElementMsg)
        rwConstellationConfig.reactionWheels = rwConfigElementList
        self.rwConstellationConfigInMsg = messaging.RWConstellationMsg().write(
            rwConstellationConfig
        )

    def _set_gateway_msgs(self) -> None:
        """Create C-wrapped gateway messages."""
        self.attRefMsg = messaging.AttRefMsg_C()
        self.attGuidMsg = messaging.AttGuidMsg_C()

        self._zero_gateway_msgs()

        # connect gateway FSW effector command msgs with the dynamics
        self.dynamics.rwStateEffector.rwMotorCmdInMsg.subscribeTo(
            self.rwNullSpace.rwMotorTorqueOutMsg
        )
        self.dynamics.thrusterSet.cmdsInMsg.subscribeTo(
            self.thrDump.thrusterOnTimeOutMsg
        )

    class MRPControlTask(fsw.SteeringImagerFSWModel.MRPControlTask):
        def _create_module_data(self) -> None:
            super()._create_module_data()

            self.rwNullSpace = self.fsw.rwNullSpace = rwNullSpace.rwNullSpace()
            self.rwNullSpace.ModelTag = "rwNullSpace"

        def _setup_fsw_objects(self, **kwargs) -> None:
            super()._setup_fsw_objects(**kwargs)
            self.set_rw_null_space(**kwargs)

        @default_args(OmegaGain=0.3)
        def set_rw_null_space(
            self,
            OmegaGain: float,
            **kwargs,
        ) -> None:
            """Define the null space to slow down the wheels."""
            self.rwNullSpace.rwMotorTorqueInMsg.subscribeTo(
                self.rwMotorTorque.rwMotorTorqueOutMsg
            )
            self.rwNullSpace.rwSpeedsInMsg.subscribeTo(
                self.fsw.dynamics.rwStateEffector.rwSpeedOutMsg
            )
            self.rwNullSpace.rwConfigInMsg.subscribeTo(
                self.fsw.rwConstellationConfigInMsg
            )
            self.rwNullSpace.OmegaGain = OmegaGain
            self._add_model_to_task(self.rwNullSpace, priority=1193)

Configure the Satellite

  • Observations:

    • SatProperties: Body angular velocity, instrument pointing direction, body position, body velocity, battery charge (properties in flight software model or dynamics model). Also, customized dynamics property in CustomDynModel above: Angle between the sun and the solar panel and four RW speed fraction.

    • OpportunityProperties: Target’s priority, normalized location, and target angle (upcoming 32 targets).

    • Time: Simulation time.

    • Eclipse: Next eclipse start and end times.

  • Actions:

    • Desat: Manage momentum for the RWs for 60 seconds.

    • Charge: Enter a sun-pointing charging mode for 60 seconds.

    • Image: Image target from upcoming 32 targets

The fault is introduced by overriding the reset_post_sim_init function. The probability of the fault occurring can be specified using the fault_chance argument, and the time of occurrence can be set using the fault_time argument.

[5]:
class CustomSatComposed(sats.ImagingSatellite):
    observation_spec = [
        obs.SatProperties(
            dict(prop="omega_BP_P", norm=0.03),
            dict(prop="c_hat_P"),
            dict(prop="r_BN_P", norm=orbitalMotion.REQ_EARTH * 1e3),
            dict(prop="v_BN_P", norm=7616.5),
            dict(prop="battery_charge_fraction"),
            dict(prop="solar_angle_norm"),
            dict(prop="wheel_speeds_frac"),
        ),
        obs.OpportunityProperties(
            dict(prop="priority"),
            dict(prop="r_LP_P", norm=orbitalMotion.REQ_EARTH * 1e3),
            dict(prop="target_angle", norm=np.pi),
            type="target",
            n_ahead_observe=32,
        ),
        obs.Time(),
        obs.Eclipse(norm=5700),
    ]

    action_spec = [
        act.Desat(duration=60.0),
        act.Charge(duration=60.0),
        act.Image(n_ahead_image=32),
    ]

    # Modified the constructor to include the fault chance and list
    def __init__(self, *args, fault_chance=0, fault_time=0.0, **kwargs) -> None:
        super().__init__(*args, **kwargs)
        self.fault_chance = fault_chance
        self.fault_time = fault_time
        self.faultList = []  # List to store faults

    def reset_post_sim_init(self) -> None:
        super().reset_post_sim_init()

        if np.random.random() < self.fault_chance:
            powerFault = RwPowerFault(
                "rwPowerLimited", self.fault_time, reducedLimit=1.0, wheelIdx=1
            )
            self.faultList = [powerFault]
            self.simulator.faultList = self.faultList
            for i in range(len(self.faultList)):
                self.faultList[i].addFaultToSimulation(self, i)

    dyn_type = CustomDynModel
    fsw_type = CustomSteeringImagerFSWModel

Configure Satellite Pareameters

When instantiating a satellite, these parameters can be overriden with a constant or rerandomized every time the environment is reset using the sat_args dictionary.

[6]:
dataStorageCapacity = 20 * 8e6 * 100
batteryStorageCapacity = 80.0 * 3600 * 2
sat_args = CustomSatComposed.default_sat_args(
    oe=random_orbit,
    imageAttErrorRequirement=0.01,
    imageRateErrorRequirement=0.01,
    batteryStorageCapacity=batteryStorageCapacity,
    storedCharge_Init=lambda: np.random.uniform(0.4, 1.0) * batteryStorageCapacity,
    u_max=0.2,  # More realistic values than 1.0
    K1=0.5,  # Updated value to have smooth and more predictable control
    nHat_B=np.array([0, 0, -1]),
    imageTargetMinimumElevation=np.radians(45),
    rwBasePower=20,
    maxWheelSpeed=1500,
    storageInit=lambda: np.random.randint(
        0 * dataStorageCapacity,
        0.01 * dataStorageCapacity,
    ),
    wheelSpeeds=lambda: np.random.uniform(-900, 900, 4),
    disturbance_vector=lambda: random_unit_vector(),
)

# Make the satellites
satellites = []
satellites.append(
    CustomSatComposed(
        "EO",
        sat_args,
        fault_chance=1.0,
        fault_time=0.0,  # Fault occurs at 0.0 (nano seconds)
    )
)

Making and interacting the Environment

For this example, the single-agent SatelliteTasking environment is used. n addition to the configured satellite, the environment requires a scenario, which defines the context in which the satellite operates. In this case, the scenario uses UniformTargets, placing 1000 uniformly distributed targets across the Earth’s surface. The environment also takes a rewarder, which defines how data collected from the scenario is rewarded. Here, UniqueImageReward is used, which assigns rewards based on the sum of the priorities of uniquely imaged targets in each episode.

[7]:
env = SatelliteTasking(
    satellite=satellites,
    terminate_on_time_limit=True,
    scenario=scene.UniformTargets(n_targets=1000),
    rewarder=data.UniqueImageReward(),
    sim_rate=0.5,
    max_step_duration=300.0,
    time_limit=95 * 60 * 3,
    log_level="INFO",
    failure_penalty=0,
    # disable_env_checker=True, # For debugging
)

First, the environment is reset. A seed is provided to ensure reproducibility of the results; it can be removed to enable randomized testing.

[8]:
observation, info = env.reset(seed=1)
2026-01-05 18:31:19,794 gym                            INFO       Resetting environment with seed=1
2026-01-05 18:31:19,796 scene.targets                  INFO       Generating 1000 targets
2026-01-05 18:31:19,892 sats.satellite.EO              INFO       <0.00> EO: Finding opportunity windows from 0.00 to 17400.00 seconds
2026-01-05 18:31:20,179 gym                            INFO       <0.00> Environment reset

The composed satellite action space returns a human-readable action map and each satellite has the same action space and similar observation space.

[9]:
print("Actions:", satellites[0].action_description)
print("States:", env.unwrapped.satellites[0].observation_description, "\n")

# Using the composed satellite features also provides a human-readable state:
for satellite in env.unwrapped.satellites:
    for k, v in satellite.observation_builder.obs_dict().items():
        print(f"{k}:  {v}")
Actions: ['action_desat', 'action_charge', 'action_image_0', 'action_image_1', 'action_image_2', 'action_image_3', 'action_image_4', 'action_image_5', 'action_image_6', 'action_image_7', 'action_image_8', 'action_image_9', 'action_image_10', 'action_image_11', 'action_image_12', 'action_image_13', 'action_image_14', 'action_image_15', 'action_image_16', 'action_image_17', 'action_image_18', 'action_image_19', 'action_image_20', 'action_image_21', 'action_image_22', 'action_image_23', 'action_image_24', 'action_image_25', 'action_image_26', 'action_image_27', 'action_image_28', 'action_image_29', 'action_image_30', 'action_image_31']
States: [np.str_('sat_props.omega_BP_P_normd[0]'), np.str_('sat_props.omega_BP_P_normd[1]'), np.str_('sat_props.omega_BP_P_normd[2]'), np.str_('sat_props.c_hat_P[0]'), np.str_('sat_props.c_hat_P[1]'), np.str_('sat_props.c_hat_P[2]'), np.str_('sat_props.r_BN_P_normd[0]'), np.str_('sat_props.r_BN_P_normd[1]'), np.str_('sat_props.r_BN_P_normd[2]'), np.str_('sat_props.v_BN_P_normd[0]'), np.str_('sat_props.v_BN_P_normd[1]'), np.str_('sat_props.v_BN_P_normd[2]'), np.str_('sat_props.battery_charge_fraction'), np.str_('sat_props.solar_angle_norm'), np.str_('sat_props.wheel_speeds_frac[0]'), np.str_('sat_props.wheel_speeds_frac[1]'), np.str_('sat_props.wheel_speeds_frac[2]'), np.str_('target.target_0.priority'), np.str_('target.target_0.r_LP_P_normd[0]'), np.str_('target.target_0.r_LP_P_normd[1]'), np.str_('target.target_0.r_LP_P_normd[2]'), np.str_('target.target_0.target_angle_normd'), np.str_('target.target_1.priority'), np.str_('target.target_1.r_LP_P_normd[0]'), np.str_('target.target_1.r_LP_P_normd[1]'), np.str_('target.target_1.r_LP_P_normd[2]'), np.str_('target.target_1.target_angle_normd'), np.str_('target.target_2.priority'), np.str_('target.target_2.r_LP_P_normd[0]'), np.str_('target.target_2.r_LP_P_normd[1]'), np.str_('target.target_2.r_LP_P_normd[2]'), np.str_('target.target_2.target_angle_normd'), np.str_('target.target_3.priority'), np.str_('target.target_3.r_LP_P_normd[0]'), np.str_('target.target_3.r_LP_P_normd[1]'), np.str_('target.target_3.r_LP_P_normd[2]'), np.str_('target.target_3.target_angle_normd'), np.str_('target.target_4.priority'), np.str_('target.target_4.r_LP_P_normd[0]'), np.str_('target.target_4.r_LP_P_normd[1]'), np.str_('target.target_4.r_LP_P_normd[2]'), np.str_('target.target_4.target_angle_normd'), np.str_('target.target_5.priority'), np.str_('target.target_5.r_LP_P_normd[0]'), np.str_('target.target_5.r_LP_P_normd[1]'), np.str_('target.target_5.r_LP_P_normd[2]'), np.str_('target.target_5.target_angle_normd'), np.str_('target.target_6.priority'), np.str_('target.target_6.r_LP_P_normd[0]'), np.str_('target.target_6.r_LP_P_normd[1]'), np.str_('target.target_6.r_LP_P_normd[2]'), np.str_('target.target_6.target_angle_normd'), np.str_('target.target_7.priority'), np.str_('target.target_7.r_LP_P_normd[0]'), np.str_('target.target_7.r_LP_P_normd[1]'), np.str_('target.target_7.r_LP_P_normd[2]'), np.str_('target.target_7.target_angle_normd'), np.str_('target.target_8.priority'), np.str_('target.target_8.r_LP_P_normd[0]'), np.str_('target.target_8.r_LP_P_normd[1]'), np.str_('target.target_8.r_LP_P_normd[2]'), np.str_('target.target_8.target_angle_normd'), np.str_('target.target_9.priority'), np.str_('target.target_9.r_LP_P_normd[0]'), np.str_('target.target_9.r_LP_P_normd[1]'), np.str_('target.target_9.r_LP_P_normd[2]'), np.str_('target.target_9.target_angle_normd'), np.str_('target.target_10.priority'), np.str_('target.target_10.r_LP_P_normd[0]'), np.str_('target.target_10.r_LP_P_normd[1]'), np.str_('target.target_10.r_LP_P_normd[2]'), np.str_('target.target_10.target_angle_normd'), np.str_('target.target_11.priority'), np.str_('target.target_11.r_LP_P_normd[0]'), np.str_('target.target_11.r_LP_P_normd[1]'), np.str_('target.target_11.r_LP_P_normd[2]'), np.str_('target.target_11.target_angle_normd'), np.str_('target.target_12.priority'), np.str_('target.target_12.r_LP_P_normd[0]'), np.str_('target.target_12.r_LP_P_normd[1]'), np.str_('target.target_12.r_LP_P_normd[2]'), np.str_('target.target_12.target_angle_normd'), np.str_('target.target_13.priority'), np.str_('target.target_13.r_LP_P_normd[0]'), np.str_('target.target_13.r_LP_P_normd[1]'), np.str_('target.target_13.r_LP_P_normd[2]'), np.str_('target.target_13.target_angle_normd'), np.str_('target.target_14.priority'), np.str_('target.target_14.r_LP_P_normd[0]'), np.str_('target.target_14.r_LP_P_normd[1]'), np.str_('target.target_14.r_LP_P_normd[2]'), np.str_('target.target_14.target_angle_normd'), np.str_('target.target_15.priority'), np.str_('target.target_15.r_LP_P_normd[0]'), np.str_('target.target_15.r_LP_P_normd[1]'), np.str_('target.target_15.r_LP_P_normd[2]'), np.str_('target.target_15.target_angle_normd'), np.str_('target.target_16.priority'), np.str_('target.target_16.r_LP_P_normd[0]'), np.str_('target.target_16.r_LP_P_normd[1]'), np.str_('target.target_16.r_LP_P_normd[2]'), np.str_('target.target_16.target_angle_normd'), np.str_('target.target_17.priority'), np.str_('target.target_17.r_LP_P_normd[0]'), np.str_('target.target_17.r_LP_P_normd[1]'), np.str_('target.target_17.r_LP_P_normd[2]'), np.str_('target.target_17.target_angle_normd'), np.str_('target.target_18.priority'), np.str_('target.target_18.r_LP_P_normd[0]'), np.str_('target.target_18.r_LP_P_normd[1]'), np.str_('target.target_18.r_LP_P_normd[2]'), np.str_('target.target_18.target_angle_normd'), np.str_('target.target_19.priority'), np.str_('target.target_19.r_LP_P_normd[0]'), np.str_('target.target_19.r_LP_P_normd[1]'), np.str_('target.target_19.r_LP_P_normd[2]'), np.str_('target.target_19.target_angle_normd'), np.str_('target.target_20.priority'), np.str_('target.target_20.r_LP_P_normd[0]'), np.str_('target.target_20.r_LP_P_normd[1]'), np.str_('target.target_20.r_LP_P_normd[2]'), np.str_('target.target_20.target_angle_normd'), np.str_('target.target_21.priority'), np.str_('target.target_21.r_LP_P_normd[0]'), np.str_('target.target_21.r_LP_P_normd[1]'), np.str_('target.target_21.r_LP_P_normd[2]'), np.str_('target.target_21.target_angle_normd'), np.str_('target.target_22.priority'), np.str_('target.target_22.r_LP_P_normd[0]'), np.str_('target.target_22.r_LP_P_normd[1]'), np.str_('target.target_22.r_LP_P_normd[2]'), np.str_('target.target_22.target_angle_normd'), np.str_('target.target_23.priority'), np.str_('target.target_23.r_LP_P_normd[0]'), np.str_('target.target_23.r_LP_P_normd[1]'), np.str_('target.target_23.r_LP_P_normd[2]'), np.str_('target.target_23.target_angle_normd'), np.str_('target.target_24.priority'), np.str_('target.target_24.r_LP_P_normd[0]'), np.str_('target.target_24.r_LP_P_normd[1]'), np.str_('target.target_24.r_LP_P_normd[2]'), np.str_('target.target_24.target_angle_normd'), np.str_('target.target_25.priority'), np.str_('target.target_25.r_LP_P_normd[0]'), np.str_('target.target_25.r_LP_P_normd[1]'), np.str_('target.target_25.r_LP_P_normd[2]'), np.str_('target.target_25.target_angle_normd'), np.str_('target.target_26.priority'), np.str_('target.target_26.r_LP_P_normd[0]'), np.str_('target.target_26.r_LP_P_normd[1]'), np.str_('target.target_26.r_LP_P_normd[2]'), np.str_('target.target_26.target_angle_normd'), np.str_('target.target_27.priority'), np.str_('target.target_27.r_LP_P_normd[0]'), np.str_('target.target_27.r_LP_P_normd[1]'), np.str_('target.target_27.r_LP_P_normd[2]'), np.str_('target.target_27.target_angle_normd'), np.str_('target.target_28.priority'), np.str_('target.target_28.r_LP_P_normd[0]'), np.str_('target.target_28.r_LP_P_normd[1]'), np.str_('target.target_28.r_LP_P_normd[2]'), np.str_('target.target_28.target_angle_normd'), np.str_('target.target_29.priority'), np.str_('target.target_29.r_LP_P_normd[0]'), np.str_('target.target_29.r_LP_P_normd[1]'), np.str_('target.target_29.r_LP_P_normd[2]'), np.str_('target.target_29.target_angle_normd'), np.str_('target.target_30.priority'), np.str_('target.target_30.r_LP_P_normd[0]'), np.str_('target.target_30.r_LP_P_normd[1]'), np.str_('target.target_30.r_LP_P_normd[2]'), np.str_('target.target_30.target_angle_normd'), np.str_('target.target_31.priority'), np.str_('target.target_31.r_LP_P_normd[0]'), np.str_('target.target_31.r_LP_P_normd[1]'), np.str_('target.target_31.r_LP_P_normd[2]'), np.str_('target.target_31.target_angle_normd'), np.str_('time'), np.str_('eclipse[0]'), np.str_('eclipse[1]')]

sat_props:  {'omega_BP_P_normd': array([0.00137284, 0.00080893, 0.00185074]), 'c_hat_P': array([-0.94095395, -0.07120216, -0.3309621 ]), 'r_BN_P_normd': array([-0.76023893, -0.76226973,  0.03873832]), 'v_BN_P_normd': array([-0.74001565,  0.72585949, -0.23976204]), 'battery_charge_fraction': 0.48805353449026784, 'solar_angle_norm': np.float64(0.3675725758375922), 'wheel_speeds_frac': array([ 0.2222634 , -0.3546573 ,  0.45374092])}
target:  {'target_0': {'priority': 0.6797657443023485, 'r_LP_P_normd': array([-0.72304393, -0.69048255,  0.02100749]), 'target_angle_normd': np.float64(0.6357597658378449)}, 'target_1': {'priority': 0.1011278274566988, 'r_LP_P_normd': array([-0.8839314 , -0.46366914, -0.06063175]), 'target_angle_normd': np.float64(0.3764086411837673)}, 'target_2': {'priority': 0.6931851990942818, 'r_LP_P_normd': array([-0.92043412, -0.36952122, -0.12749548]), 'target_angle_normd': np.float64(0.3723877146688926)}, 'target_3': {'priority': 0.17225514293500632, 'r_LP_P_normd': array([-0.92565997, -0.37833103, -0.00438879]), 'target_angle_normd': np.float64(0.38981813759316086)}, 'target_4': {'priority': 0.5711709172856598, 'r_LP_P_normd': array([-0.96343073, -0.26718398, -0.02034567]), 'target_angle_normd': np.float64(0.3943146022248573)}, 'target_5': {'priority': 0.39915339691165386, 'r_LP_P_normd': array([-0.9832796 , -0.17545716, -0.04874432]), 'target_angle_normd': np.float64(0.39937897834300234)}, 'target_6': {'priority': 0.3659991252731889, 'r_LP_P_normd': array([-0.99415125, -0.07715346, -0.07556872]), 'target_angle_normd': np.float64(0.4078843692651548)}, 'target_7': {'priority': 0.8196046614443331, 'r_LP_P_normd': array([-0.99765329, -0.01207943, -0.06739442]), 'target_angle_normd': np.float64(0.4168294389284707)}, 'target_8': {'priority': 0.31321450974410614, 'r_LP_P_normd': array([-0.97449116,  0.1725315 , -0.1435265 ]), 'target_angle_normd': np.float64(0.4358647819265422)}, 'target_9': {'priority': 0.9484757208286156, 'r_LP_P_normd': array([-0.96031054,  0.17441228, -0.21767872]), 'target_angle_normd': np.float64(0.433242760838707)}, 'target_10': {'priority': 0.48592850306846924, 'r_LP_P_normd': array([-0.96853779,  0.21402263, -0.12699944]), 'target_angle_normd': np.float64(0.4426489488945429)}, 'target_11': {'priority': 0.6184129633885858, 'r_LP_P_normd': array([-0.96210743,  0.21701823, -0.16508293]), 'target_angle_normd': np.float64(0.4411215055325246)}, 'target_12': {'priority': 0.17947175160982998, 'r_LP_P_normd': array([-0.94017117,  0.243405  , -0.23839502]), 'target_angle_normd': np.float64(0.4427323667761162)}, 'target_13': {'priority': 0.7384995398085523, 'r_LP_P_normd': array([-0.80487741,  0.52282763, -0.28075545]), 'target_angle_normd': np.float64(0.48648454206124414)}, 'target_14': {'priority': 0.27114812998614257, 'r_LP_P_normd': array([-0.76729423,  0.58576762, -0.26102845]), 'target_angle_normd': np.float64(0.4977256880591026)}, 'target_15': {'priority': 0.602211552115518, 'r_LP_P_normd': array([-0.69808315,  0.6848033 , -0.20910371]), 'target_angle_normd': np.float64(0.5172281968118192)}, 'target_16': {'priority': 0.379803286768697, 'r_LP_P_normd': array([-0.49501598,  0.82573585, -0.27040613]), 'target_angle_normd': np.float64(0.5507581988576493)}, 'target_17': {'priority': 0.4436831213331952, 'r_LP_P_normd': array([-0.45042693,  0.86931897, -0.20347018]), 'target_angle_normd': np.float64(0.5625207407345922)}, 'target_18': {'priority': 0.7048706468084478, 'r_LP_P_normd': array([-0.21420724,  0.9326702 , -0.29024395]), 'target_angle_normd': np.float64(0.59373704837557)}, 'target_19': {'priority': 0.9285111717464954, 'r_LP_P_normd': array([-0.19966509,  0.94633197, -0.25414496]), 'target_angle_normd': np.float64(0.5980174335717613)}, 'target_20': {'priority': 0.6283839193934228, 'r_LP_P_normd': array([ 0.04079828,  0.96062029, -0.27485298]), 'target_angle_normd': np.float64(0.6314125127394534)}, 'target_21': {'priority': 0.1929743249397491, 'r_LP_P_normd': array([ 0.0355612 ,  0.97627369, -0.21360027]), 'target_angle_normd': np.float64(0.6341742328673365)}, 'target_22': {'priority': 0.44341724161916973, 'r_LP_P_normd': array([ 0.11710105,  0.96569059, -0.23179522]), 'target_angle_normd': np.float64(0.6446733019141536)}, 'target_23': {'priority': 0.11836853522372437, 'r_LP_P_normd': array([ 0.34219224,  0.92570533, -0.16116486]), 'target_angle_normd': np.float64(0.6810789969475336)}, 'target_24': {'priority': 0.8270836989643272, 'r_LP_P_normd': array([ 0.37614852,  0.90638066, -0.19231846]), 'target_angle_normd': np.float64(0.6842502087329536)}, 'target_25': {'priority': 0.5185496026201819, 'r_LP_P_normd': array([ 0.38424706,  0.90105932, -0.20111263]), 'target_angle_normd': np.float64(0.6849377870859835)}, 'target_26': {'priority': 0.9675170836931263, 'r_LP_P_normd': array([ 0.38356319,  0.9111912 , -0.15036584]), 'target_angle_normd': np.float64(0.6878383017756565)}, 'target_27': {'priority': 0.9065897890064923, 'r_LP_P_normd': array([ 0.55753272,  0.82426831, -0.09868644]), 'target_angle_normd': np.float64(0.7179839245481688)}, 'target_28': {'priority': 0.9282669521531632, 'r_LP_P_normd': array([ 0.62887363,  0.76646723, -0.13056009]), 'target_angle_normd': np.float64(0.7278214487998897)}, 'target_29': {'priority': 0.07379201140065461, 'r_LP_P_normd': array([ 0.83349796,  0.55251892, -0.00199665]), 'target_angle_normd': np.float64(0.7770010755992804)}, 'target_30': {'priority': 0.010627938976362383, 'r_LP_P_normd': array([ 0.87933586,  0.46953401, -0.07941194]), 'target_angle_normd': np.float64(0.7821048162611637)}, 'target_31': {'priority': 0.13642904696262903, 'r_LP_P_normd': array([ 0.90564527,  0.41905752, -0.06478763]), 'target_angle_normd': np.float64(0.7904117166194159)}}
time:  0.0
eclipse:  [np.float64(0.7684210526315789), np.float64(0.14210526315789473)]

The simulation runs until either the battery is depleted, a RW exceeds its maximum speed (both considered failures), or a timeout occurs (which simply stops the simulation).

[10]:
total_reward = 0.0
while True:
    observation, reward, terminated, truncated, info = env.step(
        env.action_space.sample()
    )
    total_reward += reward
    if terminated or truncated:
        print("Episode complete.")
        break

print("Total reward:", total_reward)
2026-01-05 18:31:20,195 gym                            INFO       <0.00> === STARTING STEP ===
2026-01-05 18:31:20,195 sats.satellite.EO              INFO       <0.00> EO: target index 28 tasked
2026-01-05 18:31:20,196 sats.satellite.EO              INFO       <0.00> EO: Target(tgt-123) tasked for imaging
2026-01-05 18:31:20,197 sats.satellite.EO              INFO       <0.00> EO: Target(tgt-123) window enabled: 2508.8 to 2611.9
2026-01-05 18:31:20,197 sats.satellite.EO              INFO       <0.00> EO: setting timed terminal event at 2611.9
2026-01-05 18:31:20,199 sats.satellite.EO              INFO       <0.00> EO: RW Power Fault: RW1's power limit reduced to 1.0 Watts at 0.0 minutes!
2026-01-05 18:31:20,200 sats.satellite.EO              INFO       <0.50> EO: imaged Target(tgt-123)
2026-01-05 18:31:20,200 data.base                      INFO       <0.50> Total reward: {'EO': 0.9282669521531632}
2026-01-05 18:31:20,201 comm.communication             INFO       <0.50> Optimizing data communication between all pairs of satellites
2026-01-05 18:31:20,201 sats.satellite.EO              INFO       <0.50> EO: Satellite EO requires retasking
2026-01-05 18:31:20,206 gym                            INFO       <0.50> Step reward: 0.9282669521531632
2026-01-05 18:31:20,207 gym                            INFO       <0.50> === STARTING STEP ===
2026-01-05 18:31:20,207 sats.satellite.EO              INFO       <0.50> EO: target index 10 tasked
2026-01-05 18:31:20,208 sats.satellite.EO              INFO       <0.50> EO: Target(tgt-977) tasked for imaging
2026-01-05 18:31:20,209 sats.satellite.EO              INFO       <0.50> EO: Target(tgt-977) window enabled: 818.3 to 900.6
2026-01-05 18:31:20,210 sats.satellite.EO              INFO       <0.50> EO: setting timed terminal event at 900.6
2026-01-05 18:31:20,211 sats.satellite.EO              INFO       <1.00> EO: imaged Target(tgt-977)
2026-01-05 18:31:20,212 data.base                      INFO       <1.00> Total reward: {'EO': 0.48592850306846924}
2026-01-05 18:31:20,212 comm.communication             INFO       <1.00> Optimizing data communication between all pairs of satellites
2026-01-05 18:31:20,213 sats.satellite.EO              INFO       <1.00> EO: Satellite EO requires retasking
2026-01-05 18:31:20,218 gym                            INFO       <1.00> Step reward: 0.48592850306846924
2026-01-05 18:31:20,218 gym                            INFO       <1.00> === STARTING STEP ===
2026-01-05 18:31:20,219 sats.satellite.EO              INFO       <1.00> EO: target index 18 tasked
2026-01-05 18:31:20,220 sats.satellite.EO              INFO       <1.00> EO: Target(tgt-239) tasked for imaging
2026-01-05 18:31:20,221 sats.satellite.EO              INFO       <1.00> EO: Target(tgt-239) window enabled: 1758.4 to 1865.8
2026-01-05 18:31:20,221 sats.satellite.EO              INFO       <1.00> EO: setting timed terminal event at 1865.8
2026-01-05 18:31:20,222 sats.satellite.EO              INFO       <1.50> EO: imaged Target(tgt-239)
2026-01-05 18:31:20,223 data.base                      INFO       <1.50> Total reward: {'EO': 0.9285111717464954}
2026-01-05 18:31:20,224 comm.communication             INFO       <1.50> Optimizing data communication between all pairs of satellites
2026-01-05 18:31:20,224 sats.satellite.EO              INFO       <1.50> EO: Satellite EO requires retasking
2026-01-05 18:31:20,229 gym                            INFO       <1.50> Step reward: 0.9285111717464954
2026-01-05 18:31:20,230 gym                            INFO       <1.50> === STARTING STEP ===
2026-01-05 18:31:20,230 sats.satellite.EO              INFO       <1.50> EO: target index 16 tasked
2026-01-05 18:31:20,231 sats.satellite.EO              INFO       <1.50> EO: Target(tgt-530) tasked for imaging
2026-01-05 18:31:20,231 sats.satellite.EO              INFO       <1.50> EO: Target(tgt-530) window enabled: 1545.1 to 1638.7
2026-01-05 18:31:20,232 sats.satellite.EO              INFO       <1.50> EO: setting timed terminal event at 1638.7
2026-01-05 18:31:20,234 sats.satellite.EO              INFO       <2.00> EO: imaged Target(tgt-530)
2026-01-05 18:31:20,235 data.base                      INFO       <2.00> Total reward: {'EO': 0.4436831213331952}
2026-01-05 18:31:20,235 comm.communication             INFO       <2.00> Optimizing data communication between all pairs of satellites
2026-01-05 18:31:20,236 sats.satellite.EO              INFO       <2.00> EO: Satellite EO requires retasking
2026-01-05 18:31:20,240 gym                            INFO       <2.00> Step reward: 0.4436831213331952
2026-01-05 18:31:20,241 gym                            INFO       <2.00> === STARTING STEP ===
2026-01-05 18:31:20,241 sats.satellite.EO              INFO       <2.00> EO: target index 1 tasked
2026-01-05 18:31:20,242 sats.satellite.EO              INFO       <2.00> EO: Target(tgt-634) tasked for imaging
2026-01-05 18:31:20,243 sats.satellite.EO              INFO       <2.00> EO: Target(tgt-634) window enabled: 214.4 to 324.0
2026-01-05 18:31:20,243 sats.satellite.EO              INFO       <2.00> EO: setting timed terminal event at 324.0
2026-01-05 18:31:20,245 sats.satellite.EO              INFO       <2.50> EO: imaged Target(tgt-634)
2026-01-05 18:31:20,245 data.base                      INFO       <2.50> Total reward: {'EO': 0.1011278274566988}
2026-01-05 18:31:20,246 comm.communication             INFO       <2.50> Optimizing data communication between all pairs of satellites
2026-01-05 18:31:20,246 sats.satellite.EO              INFO       <2.50> EO: Satellite EO requires retasking
2026-01-05 18:31:20,251 gym                            INFO       <2.50> Step reward: 0.1011278274566988
2026-01-05 18:31:20,252 gym                            INFO       <2.50> === STARTING STEP ===
2026-01-05 18:31:20,253 sats.satellite.EO              INFO       <2.50> EO: target index 26 tasked
2026-01-05 18:31:20,254 sats.satellite.EO              INFO       <2.50> EO: Target(tgt-723) tasked for imaging
2026-01-05 18:31:20,255 sats.satellite.EO              INFO       <2.50> EO: Target(tgt-723) window enabled: 2921.5 to 2962.2
2026-01-05 18:31:20,255 sats.satellite.EO              INFO       <2.50> EO: setting timed terminal event at 2962.2
2026-01-05 18:31:20,256 sats.satellite.EO              INFO       <3.00> EO: imaged Target(tgt-723)
2026-01-05 18:31:20,257 data.base                      INFO       <3.00> Total reward: {'EO': 0.13642904696262903}
2026-01-05 18:31:20,257 comm.communication             INFO       <3.00> Optimizing data communication between all pairs of satellites
2026-01-05 18:31:20,258 sats.satellite.EO              INFO       <3.00> EO: Satellite EO requires retasking
2026-01-05 18:31:20,263 gym                            INFO       <3.00> Step reward: 0.13642904696262903
2026-01-05 18:31:20,264 gym                            INFO       <3.00> === STARTING STEP ===
2026-01-05 18:31:20,264 sats.satellite.EO              INFO       <3.00> EO: target index 31 tasked
2026-01-05 18:31:20,265 sats.satellite.EO              INFO       <3.00> EO: Target(tgt-397) tasked for imaging
2026-01-05 18:31:20,266 sats.satellite.EO              INFO       <3.00> EO: Target(tgt-397) window enabled: 3255.9 to 3347.7
2026-01-05 18:31:20,267 sats.satellite.EO              INFO       <3.00> EO: setting timed terminal event at 3347.7
2026-01-05 18:31:20,268 sats.satellite.EO              INFO       <3.50> EO: imaged Target(tgt-397)
2026-01-05 18:31:20,268 data.base                      INFO       <3.50> Total reward: {'EO': 0.7981950456811603}
2026-01-05 18:31:20,269 comm.communication             INFO       <3.50> Optimizing data communication between all pairs of satellites
2026-01-05 18:31:20,270 sats.satellite.EO              INFO       <3.50> EO: Satellite EO requires retasking
2026-01-05 18:31:20,274 gym                            INFO       <3.50> Step reward: 0.7981950456811603
2026-01-05 18:31:20,275 gym                            INFO       <3.50> === STARTING STEP ===
2026-01-05 18:31:20,276 sats.satellite.EO              INFO       <3.50> EO: target index 26 tasked
2026-01-05 18:31:20,276 sats.satellite.EO              INFO       <3.50> EO: Target(tgt-158) tasked for imaging
2026-01-05 18:31:20,277 sats.satellite.EO              INFO       <3.50> EO: Target(tgt-158) window enabled: 2956.6 to 3025.3
2026-01-05 18:31:20,278 sats.satellite.EO              INFO       <3.50> EO: setting timed terminal event at 3025.3
2026-01-05 18:31:20,279 sats.satellite.EO              INFO       <4.00> EO: imaged Target(tgt-158)
2026-01-05 18:31:20,279 data.base                      INFO       <4.00> Total reward: {'EO': 0.3900733756140301}
2026-01-05 18:31:20,280 comm.communication             INFO       <4.00> Optimizing data communication between all pairs of satellites
2026-01-05 18:31:20,281 sats.satellite.EO              INFO       <4.00> EO: Satellite EO requires retasking
2026-01-05 18:31:20,285 gym                            INFO       <4.00> Step reward: 0.3900733756140301
2026-01-05 18:31:20,286 gym                            INFO       <4.00> === STARTING STEP ===
2026-01-05 18:31:20,286 sats.satellite.EO              INFO       <4.00> EO: target index 3 tasked
2026-01-05 18:31:20,286 sats.satellite.EO              INFO       <4.00> EO: Target(tgt-706) tasked for imaging
2026-01-05 18:31:20,288 sats.satellite.EO              INFO       <4.00> EO: Target(tgt-706) window enabled: 409.7 to 464.9
2026-01-05 18:31:20,288 sats.satellite.EO              INFO       <4.00> EO: setting timed terminal event at 464.9
2026-01-05 18:31:20,289 sats.satellite.EO              INFO       <4.50> EO: imaged Target(tgt-706)
2026-01-05 18:31:20,290 data.base                      INFO       <4.50> Total reward: {'EO': 0.5711709172856598}
2026-01-05 18:31:20,291 comm.communication             INFO       <4.50> Optimizing data communication between all pairs of satellites
2026-01-05 18:31:20,291 sats.satellite.EO              INFO       <4.50> EO: Satellite EO requires retasking
2026-01-05 18:31:20,296 gym                            INFO       <4.50> Step reward: 0.5711709172856598
2026-01-05 18:31:20,296 gym                            INFO       <4.50> === STARTING STEP ===
2026-01-05 18:31:20,297 sats.satellite.EO              INFO       <4.50> EO: target index 28 tasked
2026-01-05 18:31:20,298 sats.satellite.EO              INFO       <4.50> EO: Target(tgt-841) tasked for imaging
2026-01-05 18:31:20,298 sats.satellite.EO              INFO       <4.50> EO: Target(tgt-841) window enabled: 3173.1 to 3281.6
2026-01-05 18:31:20,299 sats.satellite.EO              INFO       <4.50> EO: setting timed terminal event at 3281.6
2026-01-05 18:31:20,373 sim.simulator                  INFO       <304.50> Max step duration reached
2026-01-05 18:31:20,374 data.base                      INFO       <304.50> Total reward: {}
2026-01-05 18:31:20,375 comm.communication             INFO       <304.50> Optimizing data communication between all pairs of satellites
2026-01-05 18:31:20,379 sats.satellite.EO              WARNING    <304.50> EO: failed rw_speeds_valid check
2026-01-05 18:31:20,380 gym                            INFO       <304.50> Step reward: 0.0
2026-01-05 18:31:20,381 gym                            INFO       <304.50> Episode terminated: True
2026-01-05 18:31:20,381 gym                            INFO       <304.50> Episode truncated: False
Episode complete.
Total reward: 4.7833859613015015