Ego Planner Swarm
Module Role¶
planning/navigation/ego-planner-swarm is the ROS 2 planning submodule imported from ZJU FAST Lab. The tracked branch is ros2_version.
It belongs to the planning layer. It owns UAV local trajectory generation, obstacle avoidance, and swarm-planning validation. It does not own waypoint orchestration, behavior trees, or state monitoring.
Directory¶
Current .gitmodules entry:
Main Packages¶
| Directory | Role |
|---|---|
src/planner/plan_manage |
planner orchestration and launch entry points |
src/planner/plan_env |
local map and environment representation |
src/planner/path_searching |
path searching |
src/planner/bspline_opt |
B-spline trajectory optimization |
src/planner/traj_utils |
trajectory messages and helpers |
src/uav_simulator/* |
upstream UAV simulation and map-generation tools |
DDS Requirement¶
The upstream README recommends CycloneDDS because the default FastDDS path can show significant runtime lag.
Install:
Switch the current terminal:
Persist it in the shell:
Verify:
Common Commands¶
Launch RViz:
Single-drone simulation:
Swarm simulation:
Large-swarm simulation:
Run with optional arguments:
cd ~/venom_ws
source install/setup.bash
ros2 launch ego_planner single_run_in_sim.launch.py use_mockamap:=True use_dynamic:=False
Arguments:
| Argument | Meaning | Default |
|---|---|---|
use_mockamap |
Map generation mode. False uses Random Forest; True uses mockamap. |
False |
use_dynamic |
Whether to consider dynamics. | False |
Boundary In VNV¶
ego-planner-swarmstays underplanning/navigation/- waypoint dispatch, task switching, and global monitoring belong under
mission/ - PX4 external-pose bridging should be assembled through
driver/venom_px4_bridgeandvenom_bringup