Anona

Modular Quadrotor Autonomy Workbench

English · 中文版 README

One plant. Many planners. Fair comparison.

Watch demo on Bilibili · Anona: Modular Quadrotor Autonomy

---

Highlights

Unified plant	Self-developed ROS 2 dynamics + cascade PID + mixer
Many planners	Path A–H on one topic contract for fair comparison
Diverse maps	Process-generated forests, mazes, dense fields, narrow corridors
Mission Console	Browser / native app: maps, planners, single & multi-UAV
Acceptance	Six one-click scenarios (6/6 PASS) + formal report

---

 

---

What is Anona?

Anona is a modular quadrotor simulation workbench on Ubuntu 22.04 + ROS 2 Humble. A fixed plant (dynamics + controller) drives swappable planners under identical topics and evaluation scripts.

Layer	Role	Packages
Plant	Odometry, IMU, motor RPM	drone_dynamics, drone_controller
World	Seeded point-cloud maps	drone_map, map_adapter
Planning	Path A–H backends	drone_planner, vendors, drone_rl_planner, …
Ops	Launch, dashboard, acceptance	drone_bringup, Mission Console

Registry: PLANNERS.md · Maps: MAPS.md

ros2 launch drone_bringup planner_sim.launch.py planner:=ego map:=official_forest

---

Quick start

sudo apt update && sudo apt install -y \
  ros-humble-desktop ros-humble-tf2-ros ros-humble-pcl-conversions \
  libpcl-dev python3-matplotlib python3-numpy

source /opt/ros/humble/setup.bash
cd ~/drone_ws
export PYTHONNOUSERSITE=1
colcon build --symlink-install
source install/setup.bash

# Mission Console
bash packaging/linux/install.sh          # native
ros2 run drone_bringup dashboard         # http://127.0.0.1:8765/

---

Acceptance (6/6)

#	Scenario	Launch
1	Hover	ros2 launch drone_bringup hover.launch.py
2	Single goal	ros2 launch drone_bringup single_goal.launch.py
3	Multi-waypoint	ros2 launch drone_bringup multi_goal.launch.py
4	Static avoidance	ros2 launch drone_bringup avoidance.launch.py
5	Narrow passage	ros2 launch drone_bringup narrow_passage.launch.py
6	Stability	ros2 launch drone_bringup stability_demo.launch.py

python3 scripts/run_acceptance.py          # six scenarios (~30 min; scenario 4
                                           #   budgets one retry for planner variance)
python3 scripts/run_conformance.py --planner ego   # single-planner contract check

Scenario 4 flies the official forest at a deliberately conservative speed with a hard grid-inflation margin — its acceptance window is 420 s and the min-obstacle criterion is 0.30 m.

 

Report: report/acceptance_report.md · Benchmark: report/planner_benchmark/comparison_report.md

---

Topic contract

Topic	Role
/drone/goal	Mission goal
/map/obstacles	Obstacle cloud
/planner/local_goal, /planner/trajectory_cmd	Planner → plant
/drone/motor_rpm_cmd	Controller → dynamics
/drone/odom	State feedback

Frames: map (ENU) → base_link.

---

Citations

• Repo: https://github.com/Jungod1121/Anona-Modular-Quadrotor-Autonomy
• EGO-Planner Swarm · GCOPTER · MIGHTY · Fast-Planner · DrQ / DrQ-v2

License

MIT