Artificial Potential Field + Dijkstra Path Planning

This repository demonstrates the implementation of a hybrid Artificial Potential Field (APF) algorithm combined with Dijkstra path planning for robust path planning in a 2D environment with obstacles.

Key features include:

• Hybrid APF + Dijkstra Implementation: Combines local Artificial Potential Field navigation with global Dijkstra path planning for robust obstacle avoidance.
• Standard APF Implementation: Uses attractive forces towards the goal and repulsive forces from obstacles based on standard potential field formulas.
• Global Path Planning: Integrates Dijkstra algorithm on a discrete grid to provide global guidance and avoid local minima.
• Configurable Parameters: Allows tuning of attractive gain (xi), repulsive gain (eta), attractive threshold (sigma0), repulsive influence radius (rho0), and circumferential gain (k_circ) for escaping local minima.
• Hybrid Planning: Combines the local APF with the global Dijkstra path planner to guide the agent and mitigate local minima issues.
• Momentum: Incorporates velocity and momentum into the simulation loop for smoother movement.
• Visualization: Generates plots showing the APF force vector field, potential contour map, 3D potential surface, Dijkstra path, obstacle locations, start/goal points, and the agent's final path.

Installation

You can install this module in your project using:

pip install path/to/this/project

Or, if published to PyPI:

pip install artificial-potential-field-example

Usage

Quick Demo

Try the built-in demos to see the APF + Dijkstra module in action:

from artificial_potential_field_example import demo_basic_usage, demo_comprehensive

# Run basic demo showing core APF + Dijkstra features
demo_basic_usage()

# Run comprehensive demo showing all APF + Dijkstra features
demo_comprehensive()

Simple Usage with APFSimulator

The easiest way to use the APF + Dijkstra module is with the APFSimulator class:

from artificial_potential_field_example import APFSimulator, generate_random_obstacles

# Define start and goal
start = (0, 0)
goal = (40, 40)

# Generate random obstacles
obstacles = generate_random_obstacles(start, goal, num_obstacles=15)

# Create simulator (combines APF with Dijkstra planning)
simulator = APFSimulator(start, goal, obstacles)

# Run simulation (uses both APF forces and Dijkstra guidance)
path, success = simulator.simulate()

# Visualize results (shows both APF field and Dijkstra path)
simulator.visualize(path)

Advanced Usage with Individual Components

For more control, you can use the individual APF and Dijkstra components:

from artificial_potential_field_example import ArtificialPotentialField, compute_global_path, types

# Define obstacles and goal
obstacles = [(10, 10, 2), (20, 20, 3)]  # (x, y, radius)
goal = (40, 40)
start = (0, 0)

# Create APF instance
apf = ArtificialPotentialField(
    goal, 
    obstacles,
    xi=1.5,      # Attractive gain
    eta=1000.0,  # Repulsive gain
    sigma0=10.0, # Attractive threshold distance
    rho0=5.0,    # Repulsive influence radius
    k_circ=100.0 # Circumferential gain for escaping local minima
)

# Compute global path using Dijkstra
global_path, grid_info = compute_global_path(start, goal, obstacles)

# Compute force at a position
force = apf.compute_total_force((5, 5))
print(f"Force at (5, 5): {force}")

API Reference

• APFSimulator: Complete simulation class combining APF with Dijkstra path planning
• ArtificialPotentialField: Core APF implementation for local force computation
• compute_global_path: Dijkstra-based global path planner on discrete grid
• generate_random_obstacles: Utility to generate random obstacles for testing
• plot_field: Visualization function for APF fields, Dijkstra paths, and obstacle maps
• demo_basic_usage: Interactive demo showing basic APF + Dijkstra features
• demo_advanced_features: Demo showing parameter tuning and advanced APF + Dijkstra features
• demo_comprehensive: Complete demo covering all APF + Dijkstra functionality
• run_demo: Simple demo function that runs the main APF + Dijkstra visualization

See the code and docstrings for more advanced usage and configuration options.