Multi-UAV Path Planning and Collaboration with Dynamic Risk Field

This repository contains the implementation code and supporting data for the paper "Multi-UAV Path Planning and Collaboration Based on Dynamic Risk Field". The project develops a novel framework for multi-unmanned aerial vehicle (UAV) path planning and collaboration, utilizing a dynamic risk field model to enhance safety and efficiency in complex environments.

Overview

The research addresses the challenge of multi-UAV coordination by introducing a dynamic risk field model (Equations 1-5), which integrates real-time environmental risks, collision avoidance (Equations 7-11, 13-15), and hierarchical path planning (Equations 16-20). Key contributions include:

• A dynamic risk field to assess and mitigate threats.
• A priority-based collision resolution mechanism.
• A comparison of the proposed method with A* and potential field algorithms.
• Analysis of risk factors such as speed and turn angle.

The codebase supports simulation experiments, visualization of trajectories, and performance metrics, with outputs aligned with the paper's figures (e.g., Figures 1-4) and tables (e.g., Tables 2-3).

Installation

To set up the environment and run the code, follow these steps:

1. Clone the repository:

   git clone https://github.com/ATOI-Ming/MultiUAV_DynamicRiskField.git
   cd MultiUAV_DynamicRiskField

2. Install the required dependencies:

   pip install -r requirements.txt

Requirements

• Python 3.8+
• numpy==1.21.0
• matplotlib==3.4.0
• scipy==1.7.0
• pandas==1.3.0

Files

Source Code (src/)

• chapter1_dynamic_risk.py: Implements the dynamic risk field model (Equations 1-5), calculating risk values based on UAV states and environmental factors.
• chapter2_collaboration.py: Handles collision detection and resolution (Equations 7-11), including priority-based optimization (Equations 13-15).
• chapter3_path_planning.py: Provides global and local path planning algorithms (Equations 16-20) for UAV navigation.
• chapter4_main_experiment.py: Integrates the above modules for multi-UAV simulation, generating trajectories and logs.
• chapter5_risk_factor_analysis.py: Analyzes the impact of speed and turn angle on the risk field (related to Figures 1-3).
• chapter6_method_comparison.py: Compares the proposed method with A* and potential field methods (related to Figure 4 and Table 3).
• utils.py: Implements A* and potential field path planning algorithms for method comparison.

Paper and Configuration

• paper/MultiUAV_DynamicRiskField.pdf: The full research paper detailing the methodology and results.
• requirements.txt: Lists Python package dependencies.
• LICENSE: MIT License for open-source usage.
• .gitignore: Specifies files to ignore (e.g., .pyc, *.svg).

Data and Visualizations

• data/: Contains CSV files (e.g., drone_logs.csv) with real-time UAV experiment data (position, velocity, risk values).
• figures/: Includes SVG files (e.g., timestep_001.svg) showing algorithm display effects on the map at each step (related to Figures 1-3).
• metrics/: Contains SVG files (e.g., risk_values_010.svg, turn_angles_010.svg) with UAV performance parameters at each step (related to Table 2).
• risk_factors/: Holds SVG files (e.g., reference_no_turn.svg) visualizing speed and turn angle effects on the risk field (related to Figures 1-3).
• maps/: Includes SVG files (e.g., proposed_map_timestep_001.svg) for each step's map in the comparison experiments (related to Figure 4).
• results/: Curated SVG files from figures/ and metrics/ representing effective experimental results for the paper (related to Figures 4 and Table 3).

Usage

1. Run the main simulation:

   python src/chapter4_main_experiment.py

   Generates trajectories, logs, and initial visualizations in figures/, metrics/, and data/.

2. Analyze risk factors:

   python src/chapter5_risk_factor_analysis.py

   Produces risk field visualizations in risk_factors/.

3. Compare methods:

   python src/chapter6_method_comparison.py

   Generates comparison maps in maps/ and performance metrics.

4. Verify outputs:

   • Check generated files in data/, figures/, metrics/, risk_factors/, and maps/.
   • Use results/ for paper-ready visuals.

Outputs

• data/drone_logs.csv: Real-time UAV data for simulation validation.
• figures/timestep_XXX.svg: Snapshots of UAV trajectories and risk fields at each timestep.
• metrics/risk_values_XXX.svg, turn_angles_XXX.svg: Performance curves for risk and maneuverability.
• risk_factors/reference_no_turn.svg, acceleration_only.svg, etc.: Visualizations of risk factor impacts.
• maps/proposed_map_timestep_XXX.svg, astar_map_timestep_XXX.svg, potential_map_timestep_XXX.svg: Maps for method comparison.
• results/: Selected figures and metrics for paper inclusion.

*Note: Due to the large number of SVG files, only key frames are uploaded. *

License

This project is licensed under the MIT License, allowing free use, modification, and distribution with proper attribution.

Acknowledgments

This work was supported by the National Key Research and Development Program of the Ministry of Science and Technology of China. Special thanks to the Tsinghua University Qi Yuan Laboratory for providing computational resources and feedback.

Contact

For questions or collaborations, please open an issue or contact [[email protected]].

Future Work

• Implement dynamic threat sources (Equation 6) in chapter4_main_experiment.py.
• Add a summary chart in maps/ for method comparison.
• Enhance utils.py with parameter sensitivity analysis for A* and potential field methods.