import numpy as np
import gem_controllers as gc
import gym_electric_motor.core
[docs]
class GemController:
"""The GemController is the base for all motor controllers in the gem-control package.
A GemController consists of multiple stages that execute different control tasks like speed-control, a reference
to current set point mapping or input and output processing.
Furthermore, the GemController has got a `GemController.make` factory function that automatically designs and tunes
a classical cascaded motor controller based on classic control techniques like the proportional-integral (PI)
controller to control a gym-electric-motor environment.
"""
@property
def signals(self):
"""Input signals of the controller"""
return []
@property
def signal_names(self):
"""Signal names of the controller"""
return []
[docs]
@classmethod
def make(
cls,
env: gym_electric_motor.core.ElectricMotorEnvironment,
env_id: str,
decoupling: bool = True,
current_safety_margin: float = 0.2,
base_current_controller: str = "PI",
base_speed_controller: str = "PI",
a: int = 4,
should_plot: bool = False,
plot_references: bool = True,
block_diagram: bool = True,
save_block_diagram_as: (str, tuple) = None,
**kwargs
):
"""A factory function that generates (and parameterizes) a matching GemController for a given gym-electric-motor
environment `env`.
Args:
env(ElectricMotorEnvironment): The GEM-Environment that the controller shall be created for.
env_id(str): The corresponding environment-id to specify the concrete environment.
decoupling(bool): Flag, if a EMF-Feedforward correction stage should be used in the pi current controller.
current_safety_margin(float in [0..1]): The ratio between the maximum current set point
the reference controller generates and the absolute current limit.
base_speed_controller('PI'/'PID'/'P'/'ThreePoint'): Selection of the basic control algorithm for the
speed controller.
base_current_controller('PI'/'PID'/'P'/'ThreePoint'): Selection of the basic control algorithm for the
current controller.
a(float): Tuning parameter of the symmetrical optimum.
plot_references(bool): Flag, if the reference values of the underlying control circuits should be plotted
block_diagram(bool): Selection whether the block diagram should be displayed
save_block_diagram_as(str, tuple): Selection of whether the block diagram should be saved
Returns:
GemController: An initialized (and tuned) instance of a controller that fits to the specified environment.
"""
cls.should_plot = should_plot
control_task = gc.utils.get_control_task(env_id)
tuner_kwargs = dict()
# Initialize the current control stage
if base_current_controller == "PI":
controller = gc.PICurrentController(
env,
env_id,
base_current_controller=base_current_controller,
decoupling=decoupling,
)
tuner_kwargs["a"] = a
tuner_kwargs["plot_references"] = plot_references
elif base_current_controller == "MPC":
controller = gc.MPCCurrentController(env, env_id, **kwargs)
else:
raise NotImplementedError(f"Unsupported base_current_controller: {base_current_controller}")
tuner_kwargs["a"] = a
tuner_kwargs["plot_references"] = plot_references
if control_task in ["TC", "SC"]:
# Initilize the operation point selection
controller = gc.TorqueController(env, env_id, current_controller=controller)
tuner_kwargs["current_safety_margin"] = current_safety_margin
if control_task == "SC":
# Initilize the speed control stage
controller = gc.PISpeedController(
env,
env_id,
torque_controller=controller,
base_speed_controller=base_speed_controller,
)
# Wrap the controller with the adapter to map the inputs and outputs to the environment
if base_current_controller != "MPC":
controller = gc.GymElectricMotorAdapter(env, env_id, controller)
# Fit the controllers parameters to the environment
controller.tune(env, env_id, **tuner_kwargs)
if block_diagram:
controller.build_block_diagram(env_id, save_block_diagram_as)
return controller
@property
def stages(self):
"""Stages of the GEM Controller"""
return self._stages
def __init__(self):
self._stages = []
[docs]
def get_signal_value(self, signal_name):
"""
Get the value of a signal calling by the signal name.
Args:
signal_name(str): Name of a signal of the state
Returns:
float
"""
return self.signals[self.signal_names.index(signal_name)]
[docs]
def control(self, state, reference):
"""
Calculate the voltage reference.
Args:
state(np.array): state of the environment
reference(np.array): speed references
Returns:
np.array: reference voltage
"""
raise NotImplementedError
[docs]
def reset(self):
"""Reset all stages of the controller"""
for stage in self._stages:
stage.reset()
[docs]
def tune(self, env, env_id, **kwargs):
pass
[docs]
def control_environment(self, env, n_steps, max_episode_length=np.inf, render_env=False):
"""
Function to control an environment with the GemController.
Args:
env(ElectricMotorEnvironment): The GEM-Environment that the controller should control.
n_steps(int): Number of iteration steps.
max_episode_length(int): Maximum length of an epsiode, after which the environment and controller should be
reset.
render_env(bool): Flag, if the states of the environment should be plotted.
"""
(state, reference), _ = env.reset()
if self.block_diagram:
self.block_diagram.open()
self.reset()
current_episode_length = 0
for _ in range(n_steps): # Simulate the environment and controller for n steps
if render_env:
env.render() # Plot the states
action = self.control(state, reference) # Calculate the action
(state, reference), _, trun, done, _ = env.step(action) # Simulate one step of the environment
if done or trun or current_episode_length >= max_episode_length:
# Reset the environment and controller
(state, reference), _ = env.reset()
self.reset()
current_episode_length = 0
current_episode_length = current_episode_length + 1
if self.block_diagram:
self.block_diagram.close()