Occupancy grid

[GijsGroote]

Creation of a 2D bitmap and 3D octomap.

Option 1: From obstacles use the poses (see implementation PseudoSensor) and dimensions (see obstacle repo ) of every obstacle. Initialize the occupancy grid with free space, and loop through every obstacle in the environment, By using geometric functions update grid cells to obstacle space if the obstacle overlaps or in inside the grid cell.

A large drawback is that for every basic shape, different functions must be created such as initialise_sphere_obstacle, initilise_box_obstacle or initilise_cylinder_obstacle. This is quite annoying especially when new obstacles are added in the future. Urdf Obstacles must be estimated by a basic shape that fully intersects with the urdf obstacle.

Option 2: Using Lidar sensors on various places to create a point cloud. Converting the point cloud to an occupancy map using

octomap octomap python

A question becomes where to put the lidar sensors, many sensors are preferred in spaces which are crowded with obstacles, and were pretty much no obstacles are, less/no lider sensors are needed.

Any comments?

[maxspahn]

This could easily be integrated in a similar manner as was done in #187.