Simple OpenAI Gym environment based on PyBullet for multi-agent reinforcement learning with quadrotors
-
The default
DroneModel.CF2X
dynamics are based on Bitcraze's Crazyflie 2.x nano-quadrotor -
Everything after a
$
is entered on a terminal, everything after>>>
is passed to a Python interpreter -
Suggestions and corrections are very welcome in the form of issues and pull requests, respectively
This is the initial list of items we have to address to in order for this package to be more applicable to our project.
- Add physics modules that calculate forces and torques related to the tether
- Add the Team Blacksheep drone parameters into a new URDF file in the asset file
- Add the Team Blacksheep drone to the enumeration inside the BaseAviary class
- Simulate an IMU with and without noise to be included in the observations
- Add state estimation based on the IMU
gym-pybullet-drones |
AirSim | Flightmare | |
---|---|---|---|
Physics | PyBullet | FastPhysicsEngine/PhysX | Ad hoc/Gazebo |
Rendering | PyBullet | Unreal Engine 4 | Unity |
Language | Python | C++/C# | C++/Python |
RGB/Depth/Segm. views | Yes | Yes | Yes |
Multi-agent control | Yes | Yes | Yes |
ROS interface | ROS2/Python | ROS/C++ | ROS/C++ |
Hardware-In-The-Loop | No | Yes | No |
Fully steppable physics | Yes | No | Yes |
Aerodynamic effects | Drag, downwash, ground | Drag | Drag |
OpenAI Gym interface |
Yes | No | Yes |
RLlib MultiAgentEnv interface |
Yes | No | No |
Simulation speed-up with respect to the wall-clock when using
- 240Hz (in simulation clock) PyBullet physics for EACH drone
- AND 48Hz (in simulation clock) PID control of EACH drone
- AND nearby obstacles AND a mildly complex background (see GIFs)
- AND 24FPS (in sim. clock), 64x48 pixel capture of 6 channels (RGBA, depth, segm.) on EACH drone
Lenovo P52 (i7-8850H/Quadro P2000) | 2020 MacBook Pro (i7-1068NG7) | |
---|---|---|
Rendering | OpenGL *** | CPU-based TinyRenderer |
Single drone, no vision | 15.5x | 16.8x |
Single drone with vision | 10.8x | 1.3x |
Multi-drone (10), no vision | 2.1x | 2.3x |
Multi-drone (5) with vision | 2.5x | 0.2x |
80 drones in 4 env, no vision | 0.8x | 0.95x |
*** on Ubuntu only, uncomment the line after
self.CLIENT = p.connect(p.DIRECT)
inBaseAviary.py
Note: use
gui=False
andaggregate_phy_steps=int(SIM_HZ/CTRL_HZ)
to optimize performance
While it is easy to—consciously or not—cherry pick statistics, ~5kHz PyBullet physics (CPU-only) is faster than AirSim (1kHz) and more accurate than Flightmare's 35kHz simple single quadcopter dynamics
Exploiting parallel computation—i.e., multiple (80) drones in multiple (4) environments (see script
parallelism.sh
)—achieves PyBullet physics updates at ~20kHz
Multi-agent 6-ch. video capture at ~750kB/s with CPU rendering (
(64*48)*(4+4+2)*24*5*0.2
) is comparable to Flightmare's 240 RGB frames/s ((32*32)*3*240
)—although in more complex Unity environments—and up to an order of magnitude faster on Ubuntu, with OpenGL rendering
The repo was written using Python 3.7 on macOS 10.15 and tested on Ubuntu 18.04
Major dependencies are gym
, pybullet
,
stable-baselines3
, rllib
and ffmpeg
(for video recording only)
$ pip install gym
$ pip install pybullet
$ pip install stable-baselines3
$ pip install 'ray[rllib]'
$ brew install ffmpeg # on macOS
$ sudo apt install ffmpeg # on Linux
Using a conda
environment on macOS,
dependencies (except ffmpeg
), can be installed from file
$ cd gym-pybullet-drones/
$ conda create -n myenv --file /files/conda_req_list.txt
The repo is structured as a Gym Environment
and can be installed with pip install --editable
$ git clone https://github.com/JacopoPan/gym-pybullet-drones.git
$ cd gym-pybullet-drones/
$ pip install -e .
There are 2 basic template scripts in examples/
: fly.py
and learn.py
fly.py
runs an independent flight using PID control implemented in classDSLPIDControl
$ conda activate myenv # If using a conda environment
$ cd gym-pybullet-drones/examples/
$ python fly.py # Try 'python fly.py -h' to show the script's customizable parameters
Tip: use the GUI's sliders and button
Use GUI RPM
to override the control with interactive inputs
$ conda activate myenv # If using a conda environment
$ cd gym-pybullet-drones/examples/
$ python learn.py # Try 'python learn.py -h' to show the script's customizable parameters
Other scripts in folder examples/
are:
compare.py
which replays and compare to a trace saved inexample_trace.pkl
$ conda activate myenv # If using a conda environment
$ cd gym-pybullet-drones/examples/
$ python compare.py # Try 'python compare.py -h' to show the script's customizable parameters
downwash.py
is a flight script with only 2 drones, to test the downwash model
$ conda activate myenv # If using a conda environment
$ cd gym-pybullet-drones/examples/
$ python downwash.py # Try 'python downwash.py -h' to show the script's customizable parameters
physics.py
is an accessory script that can be used to understand PyBullet's force and torque APIs for different URDF links inp.WORLD_FRAME
andp.LINK_FRAME
$ conda activate myenv # If using a conda environment
$ cd gym-pybullet-drones/examples/
$ python physics.py # Try 'python physics.py -h' to show the script's customizable parameters
Tip: also check the examples in pybullet-examples
_dev.py
is an always in beta script with the latest features ofgym-pybullet-drones
like RGB, depth and segmentation views from each drone's POV or compatibility with RLlibs'sMultiAgentEnv
class
$ conda activate myenv # If using a conda environment
$ cd gym-pybullet-drones/examples/
$ python _dev.py # Try 'python _dev.py -h' to show the script's customizable parameters
A flight arena for one (ore more) quadrotor can be created as a child class of BaseAviary()
>>> env = BaseAviary(
>>> drone_model=DroneModel.CF2X, # See DroneModel Enum class for other quadcopter models
>>> num_drones=1, # Number of drones
>>> neighbourhood_radius=np.inf, # Distance at which drones are considered neighbors, only used for multiple drones
>>> initial_xyzs=None, # Initial XYZ positions of the drones
>>> initial_rpys=None, # Initial roll, pitch, and yaw of the drones in radians
>>> physics: Physics=Physics.PYB, # Choice of (PyBullet) physics implementation
>>> freq=240, # Stepping frequency of the simulation
>>> aggregate_phy_steps=1, # Number of physics updates within each call to BaseAviary.step()
>>> gui=True, # Whether to display PyBullet's GUI, only use this for debbuging
>>> record=False, # Whether to save a .mp4 video (if gui=True) or .png frames (if gui=False) in gym-pybullet-drones/files/, see script /files/ffmpeg_png2mp4.sh for encoding
>>> obstacles=False, # Whether to add obstacles to the environment
>>> user_debug_gui=True) # Whether to use addUserDebugLine and addUserDebugParameter calls (it can slow down the GUI)
And instantiated with gym.make()
—see learn.py
for an example
>>> env = gym.make('rl-takeoff-aviary-v0') # See learn.py
Then, the environment can be stepped with
>>> obs = env.reset()
>>> for _ in range(10*240):
>>> obs, reward, done, info = env.step(env.action_space.sample())
>>> env.render()
>>> if done: obs = env.reset()
>>> env.close()
A new environment can be created as a child class of BaseAviary
(i.e. class NewAviary(BaseAviary): ...
) and implementing the following 7 abstract methods
>>> #### 1
>>> def _actionSpace(self):
>>> # e.g. return spaces.Box(low=np.zeros(4), high=np.ones(4), dtype=np.float32)
>>> #### 2
>>> def _observationSpace(self):
>>> # e.g. return spaces.Box(low=np.zeros(20), high=np.ones(20), dtype=np.float32)
>>> #### 3
>>> def _computeObs(self):
>>> # e.g. return self._getDroneStateVector(0)
>>> #### 4
>>> def _preprocessAction(self, action):
>>> # e.g. return np.clip(action, 0, 1)
>>> #### 5
>>> def _computeReward(self, obs):
>>> # e.g. return -1
>>> #### 6
>>> def _computeDone(self, obs):
>>> # e.g. return False
>>> #### 7
>>> def _computeInfo(self, obs):
>>> # e.g. return {"answer": 42} # Calculated by the Deep Thought supercomputer in 7.5M years
See CtrlAviary
, VisionCtrlAviary
, FlockAviary
, TakeoffAviary
, and DynCtrlAviary
for examples
The action space's definition of an environment must be implemented in each child of BaseAviary
by function
>>> def _actionSpace(self):
>>> ...
In CtrlAviary
and VisionCtrlAviary
, it is a Dict()
of Box(4,)
containing the drones' commanded RPMs
The dictionary's keys are "0"
, "1"
, .., "n"
—where n
is the number of drones
The action space of FlockAviary
has the same structure but values are normalized in range [-1, 1]
The action space of TakeoffAviary
is a single Box(4,)
normalized to the [-1, 1]
range
Each child of BaseAviary
also needs to implement a preprocessing step translating actions into RPMs
>>> def _preprocessAction(self, action):
>>> ...
CtrlAviary
, VisionCtrlAviary
, FlockAviary
, and TakeoffAviary
all simply clip the inputs to MAX_RPM
DynCtrlAviary
's action
input to DynCtrlAviary.step()
is a Dict()
of Box(4,)
containing
- The desired thrust along the drone's z-axis
- The desired torque around the drone's x-axis
- The desired torque around the drone's y-axis
- The desired torque around the drone's z-axis
From these, desired RPMs are computed by DynCtrlAviary._preprocessAction()
The observation space's definition of an environment must be implemented by every child of BaseAviary
>>> def _observationSpace(self):
>>> ...
In CtrlAviary
, it is a Dict()
of pairs {"state": Box(20,), "neighbors": MultiBinary(num_drones)}
The dictionary's keys are "0"
, "1"
, .., "n"
—where n
is the number of drones
Each Box(20,)
contains the drone's
- X, Y, Z position in
WORLD_FRAME
(in meters, 3 values) - Quaternion orientation in
WORLD_FRAME
(4 values) - Roll, pitch and yaw angles in
WORLD_FRAME
(in radians, 3 values) - The velocity vector in
WORLD_FRAME
(in m/s, 3 values) - Angular velocities in
WORLD_FRAME
(in rad/s, 3 values) - Motors' speeds (in RPMs, 4 values)
Each MultiBinary(num_drones)
contains the drone's own row of the multi-robot system adjacency matrix
The observation space of FlockAviary
has the same structure but normalized to the [-1, 1]
range
The observation space of TakeoffAviary
is a single Box(20,)
, normalized to the [-1, 1]
range
The observation space of VisionCtrlAviary
is the same asCtrlAviary
but also includes keys rgb
, dep
, and seg
(in each drone's dictionary) for the matrices containing the drone's RGB, depth, and segmentation views
To fill/customize the content of each obs
, every child of BaseAviary
needs to implement
>>> def _computeObs(self, action):
>>> ...
See BaseAviary._exportImage()
) and its use in VisionCtrlAviary._computeObs()
to save frames as PNGs
Reward
, done
and info
are computed by the implementation of 3 functions in every child of BaseAviary
>>> def _computeReward(self, obs):
>>> ... # float or dict of floats
>>> def _computeDone(self, obs):
>>> ... # bool or dict of bools
>>> def _computeInfo(self, obs):
>>> ... # dict or dict of dicts
See TakeoffAviary
and FlockAviary
for example implementations
Simple drag, ground effect, and downwash models can be included in the simulation initializing BaseAviary()
with physics=Physics.PYB_GND_DRAG_DW
, these are based on the system identification of Forster (2015) (Eq. 4.2), the analytical model used as a baseline for comparison by Shi et al. (2019) (Eq. 15), and DSL's experimental work
Check the implementations of _drag()
, _groundEffect()
, and _downwash()
in BaseAviary
for more detail
Objects can be added to an environment using loadURDF
(or loadSDF
, loadMJCF
) in method _addObstacles()
>>> def _addObstacles(self):
>>> ...
>>> p.loadURDF("sphere2.urdf", [0,0,0], p.getQuaternionFromEuler([0,0,0]), physicsClientId=self.CLIENT)
Folder control
contains the implementations of 2 PID controllers
DSLPIDControl
(for DroneModel.CF2X/P
) and SimplePIDControl
(for DroneModel.HB
) can be used as
>>> ctrl = [DSLPIDControl(env) for i in range(num_drones)] # Initialize "num_drones" controllers
>>> ...
>>> for i in range(num_drones): # Compute control for each drone
>>> action[str(i)], _, _ = ctrl[i].computeControlFromState(. # Write the action in a dictionary
>>> control_timestep=env.TIMESTEP,
>>> state=obs[str(i)]["state"],
>>> target_pos=TARGET_POS)
Class Logger
contains helper functions to save and plot simulation data, as in this example
>>> logger = Logger(logging_freq_hz=freq, num_drones=num_drones) # Initialize the logger
>>> ...
>>> for i in range(NUM_DRONES): # Log information for each drone
>>> logger.log(drone=i,
>>> timestamp=K/env.SIM_FREQ,
>>> state= obs[str(i)]["state"],
>>> control=np.hstack([ TARGET_POS, np.zeros(9) ]))
>>> ...
>>> logger.save() # Save data to file
>>> logger.plot() # Plot data
Workspace ros2
contains two ROS2 Foxy Fitzroy Python nodes
AviaryWrapper
is a wrapper node for a single-droneCtrlAviary
environmentRandomControl
readsAviaryWrapper
'sobs
topic and publishes random RPMs on topicaction
With ROS2 installed (on either macOS or Ubuntu, edit ros2_and_pkg_setups.(zsh/bash)
accordingly), run
$ cd gym-pybullet-drones/ros2/
$ conda activate myenv # If using a conda environment
$ source ros2_and_pkg_setups.zsh # On macOS, on Ubuntu use $ source ros2_and_pkg_setups.bash
$ colcon build --packages-select ros2_gym_pybullet_drones
$ source ros2_and_pkg_setups.zsh # On macOS, on Ubuntu use $ source ros2_and_pkg_setups.bash
$ ros2 run ros2_gym_pybullet_drones aviary_wrapper
In a new terminal terminal, run
$ cd gym-pybullet-drones/ros2/
$ conda activate myenv # If using a conda environment
$ source ros2_and_pkg_setups.zsh # On macOS, on Ubuntu use $ source ros2_and_pkg_setups.bash
$ ros2 run ros2_gym_pybullet_drones random_control
If you wish, please cite this work as
@MISC{gym-pybullet-drones2020,
author = {Panerati, Jacopo and Zheng, Hehui and Zhou, SiQi and Xu, James and Prorok, Amanda and Sch\"{o}llig, Angela P.},
title = {Learning to Fly: a PyBullet-based Gym environment to simulate and learn the control of multiple nano-quadcopters},
howpublished = {\url{https://github.com/JacopoPan/gym-pybullet-drones}},
year = {2020}
}
- Nathan Michael, Daniel Mellinger, Quentin Lindsey, Vijay Kumar (2010) The GRASP Multiple Micro UAV Testbed
- Benoit Landry (2014) Planning and Control for Quadrotor Flight through Cluttered Environments
- Julian Forster (2015) System Identification of the Crazyflie 2.0 Nano Quadrocopter
- Carlos Luis and Jeroome Le Ny (2016) Design of a Trajectory Tracking Controller for a Nanoquadcopter
- Shital Shah, Debadeepta Dey, Chris Lovett, and Ashish Kapoor (2017) AirSim: High-Fidelity Visual and Physical Simulation for Autonomous Vehicles
- Guanya Shi, Xichen Shi, Michael O’Connell, Rose Yu, Kamyar Azizzadenesheli, Animashree Anandkumar, Yisong Yue, and Soon-Jo Chung (2019) Neural Lander: Stable Drone Landing Control Using Learned Dynamics
- Yunlong Song, Selim Naji, Elia Kaufmann, Antonio Loquercio, and Davide Scaramuzza (2020) Flightmare: A Flexible Quadrotor Simulator
University of Toronto's Dynamic Systems Lab / Vector Institute / University of Cambridge's Prorok Lab / Mitacs