Wind

Overview

PyFlyt supports various wind field models, as well as provides the capability for users to define their own wind field models.

Preimplemented Wind Models

Several popular wind models are provided:

1. Lorem Ipsum

They can be initialized in the aviary like so:

env = Aviary(..., wind_type="Lorem Ipsum")

Simple Custom Wind Modelling

For simple, stateless wind models (models that do not have state variables), it is possible to simply define the wind model as a Python method. Then, the wind model can be hooked to the aviary using the register_wind_field_function method. The following is an example:

"""Implements a simple time invariant, stateless wind model."""

import numpy as np

from PyFlyt.core import Aviary


# define the wind field
def simple_wind(time: float, position: np.ndarray):
    """Defines a simple wind updraft model.

    Args:
        time (float): time
        position (np.ndarray): position as an (n, 3) array

    """
    # the xy velocities are 0...
    wind = np.zeros_like(position)

    # and the vertical velocity is dependent on the logarithmic of height
    wind[:, -1] = np.log(position[:, -1])

    return wind


# the starting position and orientations
start_pos = np.array([[0.0, 0.0, 1.0]])
start_orn = np.array([[0.0, 0.0, 0.0]])

# environment setup, attach the windfield
env = Aviary(start_pos=start_pos, start_orn=start_orn, render=True, drone_type="quadx")
env.register_wind_field_function(simple_wind)

# set the flight mode
env.set_mode(7)

# simulate for 1000 steps (1000/120 ~= 8 seconds)
for i in range(1000):
    env.step()

env.disconnect()

More Complex Custom Wind Modelling

To define custom wind models, refer the the example provided by the WindFieldClass below:

class PyFlyt.core.abstractions.WindFieldClass(np_random: None | Generator = None)

Basic WindField class to implement custom wind field models.

Example Usage:

>>> from PyFlyt.core import Aviary
>>> from PyFlyt.core.abstractions import WindFieldClass
>>>
>>> # define the wind field
>>> class MyWindField(WindFieldClass):
>>>     def __init__(self, my_parameter=1.0, np_random: None | np.random.Generator = None):
>>>         super().__init__(np_random)
>>>         self.strength = my_parameter
>>>
>>>     def __call__(self, time: float, position: np.ndarray):
>>>         # simulate a thermal windfield, where the xy velocities are 0,
>>>         wind = np.zeros_like(position)
>>>
>>>         # but the z velocity varies to the log of height,
>>>         wind[:, -1] = np.log(position[:, -1]) * self.strength
>>>
>>>         # plus some noise,
>>>         wind += self.np_random.randn(*wind.shape)
>>>
>>>         return wind
>>>
>>> # environment setup, attach the windfield, override the parameter `my_parameter` from the default
>>> env = Aviary(..., wind_type=MyWindField, wind_options=dict(my_parameter=1.2))
>>>
>>> # step as usual
>>> ...

Default Attributes

property PyFlyt.core.abstractions.WindFieldClass.np_random

dtype - np.random.RandomState

Required Methods

PyFlyt.core.abstractions.WindFieldClass.__init__(self, np_random: None | Generator = None)

Initializes the wind_field.

PyFlyt.core.abstractions.DroneClass.__call__(*args, **kwargs)

Call self as a function.