Parameter Configuration
A simulation in SADYCOS can be customized by the user in two ways:
- by implementing the desired system behavior in the subsystem functions as explained in the page Subsystem Functions and
- by setting constant parameters of the simulation.
The latter is the topic of this page.
Page Contents
Process
After creating an object of a configuration class, it will have a property called parameters_cells, which is a cell array of structures that hold parameter values for use during the setup and run of a simulation. The values of these cell array elements must be set by the user through the output of the configuration class’ static method configureParameters which is automatically called by the configuration class’ constructor. Afterwards, the parameter_cells property is used to configure bus objects explained in the page Buses Configuration, and to create Simulink.SimulationInput objects, explained in Configuring Simulation Inputs.
Each element of the cell array will result in an individual Simulink.SimulationInput object. The run method of the configuration class will perform a simulation for every one of these objects giving the user the ability to run multiple similar simulations with differing parameters with a single function call.
Cell Array of Parameter Structures
A parameter structure holds general settings for the Simulink model itself (like the simulation duration) and constant parameters that models can use during the simulation. It must adhere to a certain format for the Simulink model to find the appropriate parameters in the right places. SADYCOS provides the class ParameterCreator in Core/Utilities to help the user create these structures. A good example of how to use this class can be found in the configureParameters method of the DefaultConfiguration class in the ExamplesMission namespace. Its code can be seen in full by opening the following code block:
configureParameters.m
function parameters_cells = configureParameters() %% General Parameters % Simulation Duration simulation_duration__s = 1000; % Simulation Mode simulation_mode = "normal"; % Pacing enable_pacing = false; pacing_rate = 1; % Send Simulation Data enable_send_sim_data = false; % Stop Criterion enable_stop_criterion = false; % Subsystem-Specific Parameters % States % Plant gravitational_parameter_Earth = 3.986004e14; InitialPlantStates.RigidBody.position_BI_I__m = [6.771e6;0;0]; InitialPlantStates.RigidBody.velocity_BI_I__m_per_s = [0;sqrt(gravitational_parameter_Earth/6.771e6);0]; InitialPlantStates.RigidBody.attitude_quaternion_BI = [1;zeros(3,1)]; InitialPlantStates.RigidBody.angular_velocity_BI_B__rad_per_s = [0;0;0]; InitialPlantStates.ReactionWheels.angular_velocities__rad_per_s = 100 * ones(3,1); % Sample Times % Sensors sensors_sample_time_parameter__s = [0.1, 0]; % GNC Algorithms gnc_algorithms_sample_time_parameter__s = [0.1, 0]; % Delays % GNC Algorithms gnc_delay = 1; InitialActuatorCommands.ReactionWheels.torque_commands__N_m = zeros(3,1); % Use helper class to prepare Parameters structure parameter_creator = ParameterCreator(InitialPlantStates, ... sensors_sample_time_parameter__s = sensors_sample_time_parameter__s, ... gnc_algorithms_sample_time_parameter__s = gnc_algorithms_sample_time_parameter__s, ... simulation_duration__s = simulation_duration__s, ... simulation_mode = simulation_mode, ... enable_pacing = enable_pacing, ... pacing_rate = pacing_rate, ... enable_send_sim_data = enable_send_sim_data, ... enable_stop_criterion = enable_stop_criterion); parameter_creator.activateDelay("GncAlgorithms", gnc_delay, InitialActuatorCommands); %% Model-Specific Parameters % Environment Models % Common mjd0 = smu.time.modifiedJulianDateFromCalDat(2025, 02, 25.5); % Time parameter_creator.addModel("Environment", ModifiedJulianDate(mjd0)) % Atmosphere parameter_creator.addModel("Environment", Nrlmsise00(mjd0, simulation_duration__s, true(24,1), 1e-2)); % Gravitational Field gravitationalField_max_Degree = 3; parameter_creator.addModel("Environment", SphericalHarmonicsGeopotential(gravitationalField_max_Degree)); % Magnetic Field magnetic_field_max_degree = 2; parameter_creator.addModel("Environment", Igrf(mjd0, simulation_duration__s, magnetic_field_max_degree)); % Plant Models % Mechanics mass__kg = 2; inertia_B_B__kg_m2 = 1*diag([1,1,1]); parameter_creator.addModel("Plant", RigidBodyMechanics(mass__kg, inertia_B_B__kg_m2)); % Aerodynamics % Get absolute path of this file's folder [this_folder,~,~] = fileparts(mfilename("fullpath")); % Get absolute paths of all obj files obj_files = string(fullfile(this_folder, 'obj_files', 'body.obj')); rotation_hinge_points_CAD = zeros(3,1); rotation_directions_CAD = [1; 0; 0]; surface_temperatures__K = num2cell(300); surface_energy_accommodation_coefficients = num2cell(0.9); DCM_B_from_CAD = [0, -1, 0;... -1, 0, 0; ... 0, 0, -1]; center_of_mass_CAD = [0; 2; 0]; parameter_creator.addModel("Plant", SimplifiedVleoAerodynamics(obj_files, ... rotation_hinge_points_CAD, ... rotation_directions_CAD, ... surface_temperatures__K, ... surface_energy_accommodation_coefficients, ... DCM_B_from_CAD, ... center_of_mass_CAD, ... false, ... 1)); % Gravity parameter_creator.addModel("Plant", PointMassGravity(mass__kg)); % Reaction Wheels reaction_wheels_spin_directions_B = eye(3); reaction_wheels_inertias__kg_m2 = 1E-4 * ones(3,1); reaction_wheels_friction_coefficients__N_m_s_per_rad = zeros(3,1); reaction_wheels_maximum_frequencies__rad_per_s = 500 * ones(3,1); parameter_creator.addModel("Plant", RateLimitedReactionWheels(reaction_wheels_inertias__kg_m2, ... reaction_wheels_spin_directions_B, ... reaction_wheels_friction_coefficients__N_m_s_per_rad, ... reaction_wheels_maximum_frequencies__rad_per_s)); % Actuator Models % Magnetic Torquers magnetic_torquers_directions = eye(3); magnetic_torquers_max_dipole_moments__A_m2 = 5 * ones(3,1); parameter_creator.addModel("Actuators", GenericMagneticTorquers(magnetic_torquers_directions, ... magnetic_torquers_max_dipole_moments__A_m2)); % GNC Algorithms Models qfr_proportional_gain = 1E-3; qfr_derivative_gain = 1E-1; parameter_creator.addModel("GncAlgorithms", QuaternionFeedbackControl(qfr_proportional_gain, qfr_derivative_gain)); %% Get Parameters Structure Parameters = parameter_creator.getParameters(); %% Add Non-Model Parameters Parameters.GncAlgorithms.reaction_wheels_spin_directions_B = reaction_wheels_spin_directions_B; Parameters.GncAlgorithms.magnetic_torquers_directions_B = magnetic_torquers_directions; Parameters.GncAlgorithms.reaction_wheels_desaturation_gain__1_per_s = 1E-2; Parameters.GncAlgorithms.reaction_wheels_inertias__kg_m2 = reaction_wheels_inertias__kg_m2; %% Write Structure to Cell for Output parameters_cells = {Parameters}; end
Roughly in the middle, an object of the ParameterCreator class is created. The constructor expects a single positional argument for the initial states of the the Plant subsystem. Besides this the user can specify a number of optional arguments for general simulation settings like the simulation duration, simulation mode, pacing, … If the user does not specify a value for these optional arguments, a default value will be used.
After the instantiation, the user can further adapt the parameters with methods. There is activateDelay to add delays to the subsystems Sensors, Actuators, and GNC Algorithms. By default, the subsystems Environment, Sensors, Actuators, and GNC Algorithms do not simulate any state dynamics. This can be changed by calling the method activateStates (which is not done in the above example). Finally, the method addModel is used to include parameters specific to models in the parameter structure. This method expects two arguments: the name of the subsystem in which the parameters should be available during simulation and an object of a model class.
After all these methods are used, the above example calls the method getParameters which assembles the parameter structure in the correct format and returns it. This structure can subsequently be amended with additional parameters that are not specific to a model before packing it into a cell and returning it. If the configuration class is supposed to run multiple simulations, the method configureParameters must return a cell array of parameter structures instead of a single cell.