Clone the repository with submodules:
git clone --recursive https://github.com/AllenIsaacJose/GO-VMP.git
Install prerequisites
- Ubuntu 20.04 LTS (Focal Fossa)
- ros-noetic-desktop-full
- Open3D v0.17
Run the setup script to configure the workspace:
cd GO-VMP
./setup_workspace.sh
Build the workspace:
catkin build
Source the workspace:
source devel_release/setup.bash
Install Docker, VS Code, and the Dev Containers extension in VS Code. The repository contains a Dockerfile and devcontainer.json in the .devcontainer folder. When opening the folder with VS Code, it should ask you whether you want to reopen the folder in a container. If not, you can press F1 and search for "Dev Containers: Rebuild and Reopen in Container". After the container is open, run "docker_setup_workspace.sh" to initialize and build the workspaces. Subsequently, you can simply run "catkin build". Run "source devel_release/setup.bash" to source the workspace.
Follow the link to install the Gurobi solver version 11.0.1 on Linux systems.
The Gurobi version used can be adjusted in view_motion_planner/CMakeLists.txt
set(GUROBI_HOME "/opt/gurobi1101/linux64")
target_link_libraries(planner_node ${catkin_LIBRARIES} ${OCTOMAP_LIBRARIES} ${YAMLCPP_LIBRARIES} ${PCL_LIBRARIES} ${OMPL_LIBRARIES} tbb gurobi_c++ gurobi110)
target_link_libraries(plan_on_map ${catkin_LIBRARIES} ${OCTOMAP_LIBRARIES} ${YAMLCPP_LIBRARIES} ${PCL_LIBRARIES} ${OMPL_LIBRARIES} tbb gurobi_c++ gurobi110)
Launch simulation:
roslaunch ur_with_cam_gazebo ur_with_cam.launch base:=trolley world_name:=world23
Run view motion planner:
rosrun view_motion_planner planner_node
Launch rviz:
roslaunch ur_with_cam_gazebo rviz.launch
Run rqt:
rqt
Select the ROI Viewpoint Planner plugin: Plugins -> Configuration -> ROI Viewpoint Planner
It should automatically switch to the view motion planner tab.
To run the complete pipeline, change "mode" to "PLAN_WITH_TROLLEY".