# HALO
**Repository Path**: Sytx_1/HALO
## Basic Information
- **Project Name**: HALO
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: MIT
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No
## Statistics
- **Stars**: 0
- **Forks**: 0
- **Created**: 2025-08-21
- **Last Updated**: 2025-08-29
## Categories & Tags
**Categories**: Uncategorized
**Tags**: None
## README
# Hierarchical Learning-Enhanced MPC for Safe Crowd Navigation with Heterogeneous Constraints
## TODO list
- [x] Release the [arXiv paper](https://arxiv.org/abs/2506.09859) in June, 2025.
- [x] Release the training and evaluation code
- [ ] Release the ros wrapper.
## Abstract
In this paper, we propose a novel **hierarchical framework** for robot navigation in dynamic environments with heterogeneous constraints. Our approach leverages a graph neural network trained via reinforcement learning (RL) to efficiently estimate the robot’s **cost-to-go**, formulated as local goal recommendations. A **spatio-temporal path-searching** module, which accounts for kinematic constraints, is then employed to generate a reference trajectory to facilitate solving the non-convex optimization problem used for explicit constraint enforcement. More importantly, we introduce an **incremental action-masking** mechanism and a **privileged learning** strategy, enabling end-to-end training of the proposed planner. Both simulation and real-world experiments demonstrate that the proposed method effectively addresses local planning in complex dynamic environments, achieving state-of-the-art (SOTA) performance. Compared with existing learning-optimization hybrid methods, our approach eliminates the dependency on high-fidelity simulation environments, offering significant advantages in computational efficiency and training scalability.
## Contributions
This paper proposes **three key techniques** that effectively address the performance limitations of traditional Model Predictive Control (MPC) in dynamic environments.
| Astar + MPC | Spatio-temporal search + MPC |
| :-------------------------------: | :-------------------------------------------: |
|  |  |
| | Better initial solution |
| **Action Mask + MPC** | **Action Mask + Privileged learning + MPC** |
|  |  |
| Better decision | Better performance and inter-frame continuity |
Real world demonstrations are available on the [Bilibili](https://www.bilibili.com/video/BV166MizgEht/?vd_source=a658930295bb4510212c5c979d654d61),[Youtube](https://www.youtube.com/watch?v=oymtbh-l1eM).
## How to Run
[Use](./assets/Run.md)
## Citation
If you find this code or paper is helpful, please kindly star :star: this repository and cite our paper by the following BibTeX entry:
```bibtex
@misc{liu2025hierarchicallearningenhancedmpcsafe,
title={Hierarchical Learning-Enhanced MPC for Safe Crowd Navigation with Heterogeneous Constraints},
author={Huajian Liu and Yixuan Feng and Wei Dong and Kunpeng Fan and Chao Wang and Yongzhuo Gao},
year={2025},
eprint={2506.09859},
archivePrefix={arXiv},
primaryClass={cs.RO},
url={https://arxiv.org/abs/2506.09859},
}
```