<title>Sensor-based planning: using a honing strategy and local map method to implement the generalized Voronoi graph</title>

Abstract

This work prescribes the procedures that are required to implement, on a conventional mobile robot, a sensor based motion planning algorithm based on the generalized Voronoi graph (GVG). The GVG is a roadmap of a static environment; recall that a roadmap is a 1D representation of an environment which the robot can use to plan a path between any two points in that environment. Once the robot has constructed the roadmap, it has in essence explored the environment. This work describes some issues in incrementally constructing the GVG with a mobile robot wit a ring of sonar sensors. Specifically, we consider some issues in specularity and dead-reckoning error reduction.