Merge dev into main branch

.github/workflows/test_solution.yml

@@ -5,6 +5,7 @@ on:
     branches:
       - master
       - main
+  workflow_dispatch:
 
 
 jobs:
@@ -25,22 +26,17 @@ jobs:
             bash
           cache-environment: true
           post-cleanup: 'all'
-      # - run: |
-      #     pip install setuptools==65.5.0 pip==21
-      #     pip install wheel==0.38.0
-      - name: Install safe-control-gym
-        run: |
-          git clone -b beta-iros-competition https://github.com/utiasDSL/safe-control-gym.git
-          cd safe-control-gym
-          pip install . --no-deps 
-          cd ..
-        shell: bash -el {0}
       - name: Install cffirmware
         run: |
           git clone https://github.com/utiasDSL/pycffirmware.git
           cd pycffirmware
           git submodule update --init --recursive
           ./wrapper/build_linux.sh
+      - name: Install stable baselines
+        run: pip install stable-baselines3
+        shell: bash -el {0}
+      - name: Install stable baselines
+        run: pip install stable-baselines3
         shell: bash -el {0}
       - run: pip install .
         shell: bash -el {0}

.gitignore

@@ -2,4 +2,8 @@
 **.egg-info/
 **/__pycache__/
 submission.csv
-logs
+logs
+secrets
+**/*.secret
+saves
+.pytest_cache

.readthedocs.yaml

@@ -0,0 +1,18 @@
+# .readthedocs.yaml
+# Read the Docs configuration file
+# See https://docs.readthedocs.io/en/stable/config-file/v2.html for details
+
+# Required
+version: 2
+
+build:
+  os: ubuntu-20.04
+  tools:
+    python: "3.8"
+
+sphinx:
+   configuration: docs/conf.py
+
+python:
+  install:
+  - requirements: docs/requirements.txt

README.md

@@ -15,6 +15,7 @@
     - [Setting up your GitHub repo for the competition](#setting-up-your-github-repo-for-the-competition)
     - [Submitting your latest iteration](#submitting-your-latest-iteration)
   - [Creating your own controller](#creating-your-own-controller)
+  - [Common errors](#common-errors)
   - [Deployment (**NOT IMPORTANT FOR STUDENTS FOR NOW**)](#deployment-not-important-for-students-for-now)
     - [Hardware setup](#hardware-setup)
     - [Common errors](#common-errors)
@@ -26,10 +27,9 @@
 
 ## Installation
 
-To run the LSY Autonomous Drone Racing project, you will need 3 main repositories:
-- [safe-control-gym](https://github.com/utiasDSL/safe-control-gym/tree/beta-iros-competition) - `beta-iros-competition` branch: The drone simulator and gym environments
+To run the LSY Autonomous Drone Racing project, you will need 2 repositories:
 - [pycffirmware](https://github.com/utiasDSL/pycffirmware) - `main` branch: A simulator for the on-board controller response of the drones we are using to accurately model their behavior
-- [lsy_drone_racing](https://github.com/utiasDSL/lsy_drone_racing) - `main` branch: This repository contains the scripts to simulate and deploy the drones in the racing challenge
+- [lsy_drone_racing](https://github.com/utiasDSL/lsy_drone_racing) - `main` branch: This repository contains the drone simulation, environments, and scripts to simulate and deploy the drones in the racing challenge
 
 ### Fork lsy_drone_racing
 
@@ -51,21 +51,14 @@ conda activate drone
 ```
 
 > **Note:** It is important you stick with **Python 3.8**. Yes, it is outdated. Yes, we'd also like to upgrade. However, there are serious issues beyond our control when deploying the code on the real drones with any other version.
-
-Next, download the `safe-control-gym` and `pycffirmware` repositories and install them. Make sure you have your conda/mamba environment active!
+Now you can install the lsy_drone_racing package in editable mode from the repository root
 
 ```bash
-cd ~/repos
-git clone -b beta-iros-competition https://github.com/utiasDSL/safe-control-gym.git
-cd safe-control-gym
-pip install .
+cd ~/repos/lsy_drone_racing
+pip install --upgrade pip
+pip install -e .
 ```
-
-> **Note:** If you receive an error installing safe-control-gym related to gym==0.21.0, run
-> ```bash
->    pip install setuptools==65.5.0 pip==21 wheel==0.38.4
-> ```
-> first
+In addition, you also need to install the pycffirmware package from source with
 
 ```bash
 cd ~/repos
@@ -78,14 +71,6 @@ conda install swig
 ./wrapper/build_linux.sh
 ```
 
-Now you can install the lsy_drone_racing package in editable mode from the repository root
-
-```bash
-cd ~/repos/lsy_drone_racing
-pip install --upgrade pip
-pip install -e .
-```
-
 Finally, you can test if the installation was successful by running 
 
 ```bash
@@ -153,6 +138,35 @@ Once you have pushed your latest iteration, a GitHub action runner will start te
 
 To implement your own controller, have a look at the [example implementation](./examples/controller.py). We recommend altering the existing example controller instead of creating your own file to not break the testing pipeline. Please also read through the documentation of the controller. You **must not** alter its function signatures. If you encounter problems implementing something with the given interface, contact one of the lecturers.
 
+## Common errors
+
+### GLIBCXX
+If you were able to install everything without any issues, but the simulation crashes when running the sim script, you should check the error messages for any errors related to `LIBGL` and `GLIBCXX_3.4.30`. If you don't find any conclusive evidence about what has happened, you might also want to run the simulation in verbose mode for `LIBGL` with
+
+```bash
+LIBGL_DEBUG=verbose python scripts/sim.py
+```
+
+Next, you should check if your system has the required library installed
+
+```bash
+strings /usr/lib/x86_64-linux-gnu/libstdc++.so.6 | grep GLIBCXX_3.4.30
+```
+
+or if it is installed in your conda environment
+
+```bash
+strings /path-to-your-conda/envs/your-env-name/lib/libstdc++.so.6 | grep GLIBCXX_3.4.30
+```
+
+If neither of those yield any results, you are missing this library and can install it with
+
+```bash
+conda install -c conda-forge gcc=12.1.0
+```
+
+If the program still crashes and complains about not finding `GLIBCXX_3.4.30`, please update your `LD_LIBRARY_PATH` variable to point to your conda environment's lib folder.
+
 ## Deployment (**NOT IMPORTANT FOR STUDENTS FOR NOW**)
 
 ### Hardware setup

benchmarks/README.md

@@ -0,0 +1,8 @@
+# Benchmarks
+
+These benchmarks are intended to test the performance of the implementation, i.e. speed and memory consumption. They are not meant to benchmark different approaches, algorithms or agents in the race task!
+
+## Profiling
+To profile the simulation, use [pyinstrument](https://github.com/joerick/pyinstrument) and run
+
+>: pyinstrument -r html profile.py

benchmarks/config/test.toml

@@ -0,0 +1,99 @@
+[sim]
+physics = "pyb"                                   # pyb: PyBullet, dyn: Mathematical dynamics model, TODO: Complete the list
+camera_view = [5.0, -40.0, -40.0, 0.5, -1.0, 0.5]
+sim_freq = 500                                    # Simulation frequency, in Hz
+ctrl_freq = 500                                   # Controller frequency, in Hz. This frequency is used to simulate the onboard controller, NOT for the environment's step function
+gui = false                                       # Enable/disable PyBullet's GUI
+
+[sim.disturbances.action]
+type = "GaussianNoise"
+std = 0.001
+
+[sim.disturbances.dynamics]
+type = "UniformNoise"
+low = [-0.1, -0.1, -0.1]
+high = [0.1, 0.1, 0.1]
+
+[env]
+reseed = false   # Whether to re-seed the random number generator between episodes
+seed = 1337      # Random seed
+freq = 30        # Frequency of the environment's step function, in Hz
+symbolic = false # Whether to include symbolic expressions in the info dict. Note: This can interfere with multiprocessing! If you want to parallelize your training, set this to false.
+
+[env.track]
+# Tall gates: 1.0m height. Short gates: 0.525m height. Height is measured from the ground to the
+# center of the gate.
+[[env.track.gates]]
+pos = [0.45, -1.0, 0.525]
+rpy = [0.0, 0.0, 2.35]
+[[env.track.gates]]
+pos = [1.0, -1.55, 1.0]
+rpy = [0.0, 0.0, -0.78]
+[[env.track.gates]]
+pos = [0.0, 0.5, 0.525]
+rpy = [0.0, 0.0, 0.0]
+[[env.track.gates]]
+pos = [-0.5, -0.5, 1.0]
+rpy = [0.0, 0.0, 3.14]
+
+# Obstacle height: 1.05m. Height is measured from the ground to the top of the obstacle.
+[[env.track.obstacles]]
+pos = [1.0, -0.5, 1.05]
+[[env.track.obstacles]]
+pos = [0.5, -1.5, 1.05]
+[[env.track.obstacles]]
+pos = [-0.5, 0.0, 1.05]
+[[env.track.obstacles]]
+pos = [0.0, 1.0, 1.05]
+
+[env.track.drone]
+pos = [1.0, 1.0, 0.05]
+rpy = [0, 0, 0]
+vel = [0, 0, 0]
+ang_vel = [0, 0, 0]
+
+[env.randomization.drone_pos]
+type = "uniform" # Everything that can be used as a distribution in numpy.random
+# Kwargs that are permissable in the np random function
+low = [-0.1, -0.1, 0.0]
+high = [0.1, 0.1, 0.02]
+
+[env.randomization.drone_rpy]
+type = "uniform"
+low = [-0.1, -0.1, -0.1]
+high = [0.1, 0.1, 0.1]
+
+[env.randomization.drone_mass]
+type = "uniform"
+low = -0.01
+high = 0.01
+
+[env.randomization.drone_inertia]
+type = "uniform"
+low = [-0.000001, -0.000001, -0.000001]
+high = [0.000001, 0.000001, 0.000001]
+
+[env.randomization.gate_pos]
+type = "uniform"
+low = [-0.1, -0.1, 0.0]
+high = [0.1, 0.1, 0.0]
+
+[env.randomization.gate_rpy]
+type = "uniform"
+low = [0.0, 0.0, -0.1]
+high = [0.0, 0.0, 0.1]
+
+[env.randomization.obstacle_pos]
+type = "uniform"
+low = [-0.1, -0.1, 0.0]
+high = [0.1, 0.1, 0.0]
+
+[[env.constraints]]
+type = "DefaultConstraint"
+ctype = "input"
+[[env.constraints]]
+type = "BoundedConstraint"
+ctype = "state"
+active_dims = [0, 2, 4]
+upper_bounds = [3, 3, 2]
+lower_bounds = [-3.0, -3.0, -0.1]

benchmarks/main.py

@@ -0,0 +1,29 @@
+from __future__ import annotations
+
+import numpy as np
+from sim import time_sim_reset, time_sim_step
+
+
+def print_benchmark_results(name: str, timings: list[float]):
+    print(f"\nResults for {name}:")
+    print(f"Mean/std: {np.mean(timings):.2e}s +- {np.std(timings):.2e}s")
+    print(f"Min time: {np.min(timings):.2e}s")
+    print(f"Max time: {np.max(timings):.2e}s")
+
+
+if __name__ == "__main__":
+    timings = time_sim_reset(n_tests=10)
+    print_benchmark_results(name="Sim reset", timings=timings)
+    sim_steps = 10
+    timings = time_sim_step(n_tests=10, sim_steps=sim_steps)
+    print_benchmark_results(name="Sim steps", timings=timings / sim_steps)
+    timings = time_sim_step(n_tests=10, sim_steps=sim_steps, physics_mode="dyn")
+    print_benchmark_results(name="Sim steps (dyn backend)", timings=timings / sim_steps)
+    timings = time_sim_step(n_tests=10, sim_steps=sim_steps, physics_mode="pyb_gnd")
+    print_benchmark_results(name="Sim steps (pyb_gnd backend)", timings=timings / sim_steps)
+    timings = time_sim_step(n_tests=10, sim_steps=sim_steps, physics_mode="pyb_drag")
+    print_benchmark_results(name="Sim steps (pyb_drag backend)", timings=timings / sim_steps)
+    timings = time_sim_step(n_tests=10, sim_steps=sim_steps, physics_mode="pyb_dw")
+    print_benchmark_results(name="Sim steps (pyb_dw backend)", timings=timings / sim_steps)
+    timings = time_sim_step(n_tests=10, sim_steps=sim_steps, physics_mode="pyb_gnd_drag_dw")
+    print_benchmark_results(name="Sim steps (pyb_gnd_drag_dw backend)", timings=timings / sim_steps)

benchmarks/profile.py

@@ -0,0 +1,19 @@
+from pathlib import Path
+
+import gymnasium
+
+import lsy_drone_racing  # noqa: F401, required for gymnasium.make
+from lsy_drone_racing.utils import load_config
+
+
+def main():
+    config = load_config(Path(__file__).parent / "config/test.toml")
+    env = gymnasium.make("DroneRacing-v0", config=config)
+    env.reset()
+    for _ in range(1000):
+        env.step(env.action_space.sample())
+    # 0.455
+
+
+if __name__ == "__main__":
+    main()

benchmarks/sim.py

@@ -0,0 +1,44 @@
+from __future__ import annotations
+
+import timeit
+from pathlib import Path
+from typing import TYPE_CHECKING
+
+import numpy as np
+
+if TYPE_CHECKING:
+    import numpy.typing as npt
+
+load_config_code = f"""
+from pathlib import Path
+
+from lsy_drone_racing.utils import load_config
+
+config = load_config(Path('{Path(__file__).parent / "config/test.toml"}'))
+
+"""
+env_setup_code = """
+import gymnasium
+
+import lsy_drone_racing
+
+env = gymnasium.make('DroneRacing-v0', config=config)
+
+"""
+
+
+def time_sim_reset(n_tests: int = 10) -> npt.NDArray[np.floating]:
+    setup = load_config_code + env_setup_code
+    stmt = """env.reset()"""
+    return np.array(timeit.repeat(stmt=stmt, setup=setup, number=1, repeat=n_tests))
+
+
+def time_sim_step(
+    n_tests: int = 10, sim_steps: int = 100, physics_mode: str = "pyb"
+) -> npt.NDArray[np.floating]:
+    modify_config_code = f"""config.sim.physics = '{physics_mode}'\n"""
+    setup = load_config_code + modify_config_code + env_setup_code + "\nenv.reset()"
+    stmt = f"""
+for _ in range({sim_steps}):
+    env.step(env.action_space.sample())"""
+    return np.array(timeit.repeat(stmt=stmt, setup=setup, number=1, repeat=n_tests))