Path planning, the art of finding a way from point ’A’ to point ’B’ is well researched in the field of robotics, but there are still challenges to address. In this thesis you will work on an addon of the path planner called D*+, to address the long computation times required to build a large map to plan on.

D*+ plans a safe detailed path that if followed is avoiding obstacles. This path will be further referred as a “local path”. The problem addressed in this proposal is as follows. In long‐term exploration missions building a map to plan on takes too much time with the increasing map size, therefore there is an idea to create a map segmentation solution that will split big map in multiple segments or “local maps”. Thus, it will be needed only to feed “local map” to D*+ for fast local path planning. And planning between the map segments will be based on a “global planner” that will make a high level “global path” thus allowing the system to plan a global path between any two points in the global map while keeping the “local path” safe.

The tasks are:

1. to create a large‐scale map segmentation solution for D*+.

2. to create a “global planner” that will plan the path between the map segments.