The term planning scene is used to represent the world around the robot and is used to store the state of the robot itself. The planning scene monitor inside of move_group maintains the planning scene representation. The move_group node consists of another section called the world geometry monitor, which builds the world geometry from the sensors of the robot and from the user input.

The planning scene monitor reads the joint_state topics from the robot, and the sensor information and world geometry from the world geometry monitor. The world scene monitor reads from the occupancy map monitor, which uses 3D perception to build a 3D representation of the environment, called Octomap. The Octomap can be generated from PointClouds, using PointCloud occupancy map update plugin. Similarly, the depth images are generated by a depth image occupancy map updater. You will see this part in the next section when we introduce perception.