kumar-lab-motion-planning-entry-level-proj

reference repo:

• https://github.com/KumarRobotics/kr_mav_control
• https://github.com/sikang/DecompUtil

About

• motion planning entry level project
• trajectory generation using optimization-based method(minimum snap)

methods

method 1

• front-end path search: A* + Ramer–Douglas–Peucker
• back-end optimization: 7th polynomial trajectory generation in dense and small corridors (reference: Minimum Snap Trajectory Generation and Control for Quadrotors)
• performance: average running time: frontend: ~30ms, backend: ~80ms with slow velocity, acceleration and jerk.
• video1

method 2

• front-end path search: A* + Ramer–Douglas–Peucker
• back-end optimization: 7th polynomial trajectory generation in convex polyhedron corridors (reference: Planning_Dynamically_Feasible_Trajectories_for_Quadrotors_Using_Safe_Flight_Corridors_in_3D_Complex_Environments)
• performance: average running time: frontend: ~30ms, backend: ~300ms with fast velocity, acceleration and jerk.
• video2

build on ROS-noetic

upzip src1.zip(method 1) or src2.zip(method 2) and rename it src

mkdir ws
catkin_make

run

• in ws path

source devel/setup.bash
roslaunch param_env structure_map.launch
rosrun kr_trackers waypoints_to_action.py __ns:=quadrotor
rosrun rqt_mav_manager rqt_mav_manager