AMSwarm 2.0 is a high-speed drone swarm trajectory planner that improves upon the original AMSwarm implementation (and associated paper).
The core of AMSwarm 2.0 is written in C++, using the Eigen linear algebra library for high-speed, real-time suitable code. We also provide an easy to use Python wrapper that allows for seamless integration to existing simulators such as gym-pybullet-drones.
- Temporal Waypoint Tracking: Targets precise arrival times at waypoints, optimizing for temporal accuracy.
- Dynamics Model Incorporation: Uses identified dynamics models for enhanced trajectory prediction and collision avoidance.
- Optimized AM Algorithm: Improves the Alternating-Minimization algorithm for greater efficiency, completely eliminating slow trigonometric operations.
- Distributed Avoidance Responsibilities: Distributes avoidance tasks among drones to reduce collision constraints by half.
- Improved Code Structure: Abstracts the AM algorithm for easier adaptation and customization to various use cases.
Tested on Ubuntu 20.04 with Python 3.8. Requirements include:
- CMake (>=3.12)
- C++ compiler with OpenMP support
- pybind11 (Note: depending on your system, you may need to install the pybind11-global option for CMake visibility)
- Clone the AMSwarm repository:
git clone https://github.com/bsprenger/AMSwarm.git
- Navigate to the cloned directory and install with pip:
pip install .
This will automatically compile the C++ code and install the Python module in the appropriate location for your Python installation.
You can use AMSwarm in C++ by including the appropriate headers in your project and linking against the compiled library. For Python usage, after installation, import the amswarm module in your scripts:
import amswarm
Refer to the examples provided in the examples/python directory for comprehensive usage scenarios.
We welcome contributions to AMSwarm! If you have suggestions for improvements or encounter any issues, please open an issue or pull request on our GitHub repository.
AMSwarm is released under the MIT License. See the LICENSE file for more details.
[1] Vivek K. Adajania, Siqi Zhou, Arun Kumar Singh, and Angela P. Schoellig. Amswarm: An alternating minimization approach for safe motion planning of quadrotor swarms in cluttered environments. In 2023 IEEE International Conference on Robotics and Automation (ICRA), pages 1421–1427, 2023.