src.environment.problem.MOO.UAV.utils

Module Contents

Functions

createmodel	Generates a simulation model for a drone navigation task in a terrain with threats.
generate_terrain	Generates a series of points that approximate terrain using a simple algorithm with minimal parameters.
interpolate	Introduction Performs scattered 2D linear interpolation from known (x0, y0, v0) to new points (xn, yn). Adds artificial boundary points to ensure smooth interpolation near edges.

API

src.environment.problem.MOO.UAV.utils.createmodel(map_size=900, r=100, rr=10, num_threats=5, rng=None)

Generates a simulation model for a drone navigation task in a terrain with threats.

This function creates a synthetic terrain map and places cylindrical threats randomly. It also determines suitable start and end points for the drone, ensuring they are not in immediate collision with any threats.

Parameters:

map_size : int, optional (default=900) Size of the square terrain grid. r : float, optional (default=100) Initial roughness of the terrain. rr : float, optional (default=10) Roughness variation factor controlling terrain detail. num_threats : int, optional (default=5) Number of cylindrical threats to be placed on the map. seed : int, optional (default=3849) Random seed for reproducibility.

Returns:

model : dict A dictionary containing terrain, threat information, and start/end points:

- 'drone_size' : float
  The size of the drone.
- 'danger_dist' : float
  Minimum safety distance from threats.
- 'threats' : ndarray, shape (num_threats, 4)
  Array where each row represents a threat: `[x, y, height, radius]`.
- 'start' : ndarray, shape (3, 1)
  Start position `[x, y, z]` for the drone.
- 'end' : ndarray, shape (3, 1)
  End position `[x, y, z]` for the drone.
- 'xmin', 'xmax', 'ymin', 'ymax', 'zmin', 'zmax' : float
  Boundaries of the map.
- 'MAPSIZE_X', 'MAPSIZE_Y' : float
  Dimensions of the terrain grid.
- 'X', 'Y' : ndarray
  Meshgrid representing terrain coordinates.
- 'H' : ndarray
  Heightmap of the terrain.

Notes:

• The threats are randomly placed with constraints ensuring they remain within bounds.

• Start and end points are selected to avoid initial collisions with threats.

• If no valid start/end points are found after 20 attempts, an assertion error is raised.

Example:

src.environment.problem.MOO.UAV.utils.generate_terrain(n=8, map_size=900, h0=130, r0=None, rr=None, rng=None)

Generates a series of points that approximate terrain using a simple algorithm with minimal parameters.

Parameters:

n : int Number of iterations of the algorithm, controlling the level of detail. Values beyond 8 add negligible visible detail but significantly increase computation time. mesh_size : int The size of the output mesh (e.g., 512 for a 512x512 grid). h0 : float Initial elevation. r0 : float Initial roughness, determining how much terrain can vary in a step. rr : float Roughness roughness, controlling how much roughness itself can vary in a step.

Returns:

hm : ndarray 2D mesh grids useful for surface plotting.

src.environment.problem.MOO.UAV.utils.interpolate(x0, y0, v0, xn, yn)

Introduction

• Performs scattered 2D linear interpolation from known (x0, y0, v0) to new points (xn, yn).

• Adds artificial boundary points to ensure smooth interpolation near edges.

Args

• x0 (ndarray): Known x-coordinates.

• y0 (ndarray): Known y-coordinates.

• v0 (ndarray): Known values at (x0, y0).

• xn (ndarray): New x-coordinates to interpolate.

• yn (ndarray): New y-coordinates to interpolate.

Returns

• v (ndarray): Interpolated values at coordinates (xn, yn), shape matches xn.