Robot map exploration

Map exploration tries to answer the question fo where I haven’t been, given a map, and how to cover an unexplored area.

A basic behavioral approach is to do this randomly, or avoid explored areas to get into new ones. These are valid techniques, but of course inefficient ones. Another behavioral approach may be to use information of an occupancy grid 20220228202048 to define new goals, but although easy to implement, this can also be innefficient.

Fo this, two basic techniques have been developed that work very well on indoor environments:

• Frontier based exploration.
• Generalized Voronoi graphs.

Frontier based exploration

This techniques uses bayesian-dampster grids fro creating the map. When a robot enters a new area, there is a boundary between each area sensed and areas that are open and unsensed. These form lines called frontiers, and these fronteiers can be used to select, from many possible frontiers, the next one to explore, such as the closest one or the biggest one (calculating the centroids as the new goal)

General voroni graphs

When moving and using Voronoi Graphs for representing the world 20220209184858, the robot generates and follows path, but when presented with an open area or a dead end, multiple paths are created. Since on dead ends we know that this graphs end at objects, we mark those as not to follow, and we can backtrack and search for the open ones as the next to explore

Notes References

20210714190242 Robotics Basics - Navigation

20210514183815 INDEX - Robotics

References

(Murphy 2000)

Murphy, Robin. 2000. Introduction to AI Robotics. Intelligent Robotics and Autonomous Agents. Cambridge, Mass: MIT Press.