TPRFly: LLM-based Third-Party Risk-Aware Decision-Making for UAVs in Dynamic Traffic Environments

arXiv | paper | website
This repository is the official implementation of TPRFly. TPRFly is a third-party risk-aware decision-making framework for UAVs operating in complex urban environments characterized by dense traffic flows and buildings.

Table of Contents

• Setup
• Datasets
• Experiments
• Acknowledgements

Setup

Clone the repository and set up the environment:

git clone https://github.com/SAS-HKU/TPRFly.git
cd TPRfly
conda create --name tprfly python==3.9
conda activate tprfly
pip install -r requirement.txt

Note: For compatibility, you may experiment with different pytorch versions. We use PyTorch 1.12.1 with cu113.

Datasets

Experiments

TODO LIST

• Introduction & DEMO
• Datasets
• Code
• Visualizations

Acknowledgements

Citation