VOID: VO-Integrated Deep Reinforcement Learning for 3D Dynamic Obstacle Avoidance

Release plan

We plan to release the full implementation once the paper is published.

Overview

This repository hosts materials related to our paper:

VOID: VO-Integrated Deep Reinforcement Learning for 3D Dynamic Obstacle Avoidance

VOID integrates 3D Velocity Obstacle (VO) constraints with deep reinforcement learning for safe navigation in dynamic environments.

Figures

Framework

The following figure illustrates the overall framework used in the paper:

Real-world experiments

The following figure shows examples from our real-world experiments:

Contents (current)

• framework.png: framework overview figure
• realworld.png: real-world experiment figure

More materials will be added progressively. The complete codebase will be released after publication.

Citation

If you find this work useful, please cite our paper. The citation information will be updated after publication.

Contact

If you have questions, feel free to open an issue in this repository.