UAV-develop/PX4-Autopilot
1
0
Fork 0
mirror of https://github.com/PX4/PX4-Autopilot.git synced 2026-07-26 16:28:06 +08:00
Code Issues Packages Projects Releases Wiki Activity

feat(sim): add RotorPy simulator support (#27116)

Browse Source 
* Add rotorpy simulator and integrate with main simulator script

* docs: add rotorpy simulator instructions

* docs(update): subedit

* docs: address RotorPy review comments

---------

Co-authored-by: Davide Iafrate <dvde.iafrate98@gmail.com>
Co-authored-by: Ramon Roche <mrpollo@gmail.com>
Co-authored-by: Hamish Willee <hamishwillee@gmail.com>
This commit is contained in:
Davide Iafrate
2026-06-26 03:51:55 +05:00
committed by GitHub
parent 3db36cb022
commit 6ddf60c7b4
6 changed files with 137 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
docs/en/SUMMARY.md
View File

@@ -889,6 +889,7 @@
- [Multi-Vehicle Sim with JMAVSim](sim_jmavsim/multi_vehicle.md)
- [JSBSim Simulation](sim_jsbsim/index.md)
- [AirSim Simulation](sim_airsim/index.md)
- [RotorPy Simulation](sim_rotorpy/index.md)
- [Hardware Simulation](simulation/hardware.md)
- [HITL Simulation](simulation/hitl.md)
- [SIH on Hardware](sim_sih/hardware.md)

96
docs/en/sim_rotorpy/index.md Normal file
View File

@@ -0,0 +1,96 @@
# RotorPy Simulation
:::warning
This simulator is [community supported and maintained](../simulation/community_supported_simulators.md).
It may or may not work with current versions of PX4.
See [Toolchain Installation](../dev_setup/dev_env.md) for information about the environments and tools supported by the core development team.
:::
RotorPy is a Python-based multirotor simulation environment with [aerodynamic wrenches](https://arxiv.org/abs/2306.04485), useful for education and research in estimation, planning, and control for UAVs.
It provides stand-alone classes and a [Gymnasium environment](https://gymnasium.farama.org/).
<lite-youtube videoid="L8-QZgc6Vwk" title="Demo of PX4 with RotorPy simulator"/>
## Why Use RotorPy?
RotorPy simulates the aerodynamic forces and moments acting on a multirotor, along with actuator limits, sensor noise, wind, and obstacles.
This makes it useful when you want to test estimation, planning, or control algorithms against more realistic vehicle dynamics than a simple point-mass or kinematic simulation.
The Gymnasium interface is useful for reinforcement learning and other Python-based research workflows.
It lets you connect policies and learning pipelines to a PX4-controlled multirotor simulation without rewriting them around PX4-specific APIs.
## Installation
RotorPy can be installed using `pip`:
```sh
pip install rotorpy[px4]
```
To install other tagged versions, see [`pyproject.toml`](https://github.com/spencerfolk/rotorpy/blob/main/pyproject.toml).
## Running the Simulation
This example will allow you to control the quadrotor through QGroundControl.
::: tip
If instead you want RotorPy to take control of the drone you can pass the `autopilot_controller=False` argument to the `PX4Multirotor` constructor when instantiating it.
This can be useful to test perception pipelines, since the environment is controlled through RotorPy's own controllers.
:::
First you need to build the SITL binary by setting up the toolchain and running:
```sh
make px4_sitl
```
Then run the PX4 SITL binary with the command:
```sh
PX4SIMULATOR=rotorpy PX4_SYS_AUTOSTART=10040 ./build/px4_sitl_default/bin/px4
```
Start the `example/px4_basic_example.py`, available [here](https://github.com/spencerfolk/rotorpy/blob/main/examples/basic_usage_px4.py) and referenced in the code below:
```python
# test_px4_sitl.py
from rotorpy.environments import Environment
from rotorpy.trajectories.circular_traj import ThreeDCircularTraj
from rotorpy.vehicles.px4_multirotor import PX4Multirotor
from rotorpy.vehicles.px4_sihsim_quadx_params import quad_params as sihsim_quadx
from rotorpy.controllers.quadrotor_control import SE3Control
from rotorpy.trajectories.hover_traj import HoverTraj
import numpy as np
circular_trajectory = ThreeDCircularTraj(radius=np.array([1,1,0]))
hover_trajectory = HoverTraj(x0=np.array([0, 0, 5]))
def main():
vehicle = PX4Multirotor(sihsim_quadx, enable_ground=True)
controller = SE3Control(sihsim_quadx)
env = Environment(
vehicle = vehicle,
controller = controller,
trajectory = circular_trajectory,
sim_rate = 100,
)
results = env.run(
t_final = 60,
use_mocap=False,
plot_mocap=False,
plot_estimator=False,
plot_imu=False,
plot = True,
animate_bool = False,
verbose = True,
)
print("Done—PX4 SITL ran for", len(results["time"]), "steps")
if __name__ == '__main__':
main()
```

1
docs/en/simulation/community_supported_simulators.md
View File

@@ -18,6 +18,7 @@ Questions about these tools should be raised on the [discussion forums](../contr
| [JMAVSim](../sim_jmavsim/index.md) | <p>A simple multirotor/quad simulator. This was previously part of the PX4 development toolchain but was removed in favour of [Gazebo](../sim_gazebo_gz/index.md).</p> <p><strong>Supported Vehicles:</strong> Quad</p> |
| [JSBSim](../sim_jsbsim/index.md) | <p>A simulator that provides advanced flight dynamics models. This can be used to model realistic flight dynamics based on wind tunnel data.</p> <p><strong>Supported Vehicles:</strong> Plane, Quad, Hex</p> |
| [AirSim](../sim_airsim/index.md) | <p>A cross platform simulator that provides physically and visually realistic simulations. This simulator is resource intensive, and requires a significantly more powerful computer than the other simulators described here.</p><p><strong>Supported Vehicles:</strong> Iris (MultiRotor model and a configuration for PX4 QuadRotor in the X configuration).</p> |
| [RotorPy](../sim_rotorpy/index.md) | <p>A Python-based multirotor simulation environment with aerodynamic forces and moments, actuator limits, sensor noise, wind, obstacles, and a Gymnasium interface for research workflows.</p><p><strong>Supported Vehicles:</strong> Quad</p> |
:::tip
[Gazebo](../sim_gazebo_gz/index.md) and [SIH](../sim_sih/index.md) are the officially supported simulators. See the [Simulation](index.md) page for more information.