lanternw / fastplanner-lite

A lite version of Fast-planner, global planner, using MINCO back-end. 代码简洁，供参考学习。

FastPlanner-Lite

后端使用minco的FastPlanner全局规划器。代码简洁，供参考学习。

更新Swarm功能。

依赖

Eigen 3
PCL 1.7
pygame (可选)

build and run

catkin_make -DCMAKE_BUILD_TYPE=Release

如果报错找不到Polynome.h 或PositionCommand.h，大概率是编译顺序问题，可以手动把Polynome.h文件放进devel/include/planner_manager里, PositionCommand.h 放进devel/include/quadrotor_msgs里。

source devel/setup.bash

roslaunch planner_manager run_in_sim.launch

使用3D Nav Goal选择目标点。 修改run_in_sim.launch的uav_count参数，可以修改接入规划的飞机数量。

map1较大，如果电脑性能弱，run_in_sim.launch 172行map_id改为2试试，这幅地图更小。

ESDF可视化工具

rosrun debug grad_viewer.py

W键和S键选择绘制的z轴高度，B键绘制，Q键退出。

fastplanner-lite's Issues

请问仿真界面的path是A*寻找的路径吗？怎么显示优化后的轨迹呢？

Discussion: most appropriate/efficient container for openset

Why is std::multimap the choice of container for openset instead of priority_queue? I realize that for both containers, if the node is already in the openset and we decide to update its parent node and total cost, the openset will not really be updated since there is no shuffling take place. Maybe this is not so problematic for actual implementation, but if we want to implement A* in its original flavour, what would be the best container for openset?