Radar-Like Multimodal UAV Swarm Simulation

This repository contains the final research implementation of a highly realistic, multimodal UAV swarm simulation. It is designed to evaluate decentralized swarm decision-making under severe perception uncertainty, including sensor noise, clutter, false alarms, radar dropouts, and communication outages.

Features

• Strict Ground-Truth Isolation: UAV runtime controllers never have access to ground truth; decisions are made exclusively on noisy, fused sensor data.
• Multimodal Perception Modeling: Simulates Radar, Vision, and LiDAR characteristics including range-dependent noise, Probability of Detection (P_D), Probability of False Alarm (P_FA), and Poisson clutter.
• Advanced Tracking & Fusion: Implements Kalman filtering with nearest-neighbor data association, track coasting, and dynamic trust-weighted multimodal fusion.
• Decentralized Communication: Simulates packet loss, latency, and complete network outages between swarm agents.
• Comprehensive Evaluation: Automated Monte Carlo experiment framework for generating rigorous statistical evidence.

Installation

Ensure you have Python 3.8+ installed.

# Clone the repository
git clone https://github.com/yourusername/uav-research.git
cd uav-research

# Install dependencies
pip install -r simulation_prototype/requirements.txt

Quick Start: Interactive Demo

To launch the interactive GUI, tweak sensor parameters in real-time, and watch a single simulation trial unfold (2D or 3D):

python simulation_prototype/run_interactive_demo.py

Reproducibility: Automated Demo

To run a small, fast, representative slice of the core experiments and generate a subset of the CSV logs (proving the pipeline works on your machine):

python simulation_prototype/run_final_demo.py

Results will be saved in simulation_prototype/results/demo/.

Running the Full Experiment Matrix

To re-run the massive batch of final experiments (hundreds of trials across all scenarios) that generated the statistics used in the final report:

python simulation_prototype/run_final_simulations.py

Warning: This process is computationally intensive and may take significant time to complete.

Project Structure

• simulation_prototype/: The core python simulation codebase, models, and execution scripts.
  • validation/: Isolated validation tests for radar, tracking, fusion, and trust models.
  • results/final_release/: The aggregated CSV metrics and statistical JSON analysis.
  • plots/final_release/: High-quality plots comparing fusion modes, latency, and collision risks.
• media/: Pre-rendered 3D videos of major scenarios (baseline, heavy clutter, communication outage, etc.).
• final_acceptance/: Contains the final technical report (Final_Report.docx), the presentation (Radar_Swarm_Simulation_Final_Presentation.pptx), and the demonstration script.
• notes/: A comprehensive collection of research literature reviews, paper classifications, methodology drafts, and simulation architecture documentation. Includes the core literature_matrix.csv and literature_summary.md.
• tables/: Planning documentation and expected results schemas.

Final Release Artifacts

The formal research findings are documented in the final_acceptance directory. Please refer to the Final Report for detailed methodology, statistical significance testing, and architectural conclusions. The Demonstration Script (final_research_demo_script.md) can be used to guide live presentations of this system.

---

This research simulation is frozen and validated against strict completion gates. Ground-truth leakage has been fully audited and removed.