Planning collision-free paths in a three-dimensional partially known environment

An algorithm that finds collision-free paths for a robot which works in three-dimensional (3D) initially unknown environments is described. Vision and proximity sensors are used in order to increase the knowledge of the environment. These sensors recognize and calculate the position and orientation of obstacles in the robot's field of view. The environment model is based on a polyhedral representation. The environment map is continuously updated and increased using sensor information. This map is based on 3D free-ways formed by pyramids with a convex base. The environment knowledge database is used for rapid collision-free path planning.